Chromium

[Empty comment from Monorail migration]

[Monorail components: Blink>Network]

What test are failing?

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/5c2514f6082f00006ab0daa25cccaffb29e3843a

commit 5c2514f6082f00006ab0daa25cccaffb29e3843a
Author: Kenichi Ishibashi <bashi@chromium.org>
Date: Mon Mar 13 06:41:47 2023

Remove quic-transport scheme

`quic-transport` scheme was introduced for the QuicTransport API, but
it was replaced with the WebTransport API. The WebTransport API uses `https` scheme. Chromium doesn't use `quic-transport` scheme in any
place.

Note that this scheme does nothing for a proxy that uses QUIC as the
underlying transport layer. A proxy may use HTTP/3 but the scheme
should be `https` in that case.

Bug: 1416006
Change-Id: I595f42e2676dc712c1aab1426de3f7c51ad61c73
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4328307
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Reviewed-by: Adam Rice <ricea@chromium.org>
Reviewed-by: Hidehiko Abe <hidehiko@chromium.org>
Reviewed-by: Charlie Harrison <csharrison@chromium.org>
Reviewed-by: Varun Khaneja <vakh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1116250}

[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/url/url_util.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/chrome/browser/ash/crosapi/network_settings_translation_unittest.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/components/safe_browsing/core/browser/utils/scheme_logger.h
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/net/cookies/cookie_constants.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/url/url_constants.h
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/net/base/url_util_unittest.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/chrome/browser/ash/crosapi/translate_crosapi_to_proxy_config.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/net/cookies/cookie_constants.h
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/url/url_constants.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/url/url_util_unittest.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/net/cookies/cookie_monster_unittest.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/services/network/public/cpp/is_potentially_trustworthy_unittest.h
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/components/safe_browsing/core/browser/utils/scheme_logger_unittest.cc
[modify] https://crrev.com/5c2514f6082f00006ab0daa25cccaffb29e3843a/components/safe_browsing/core/browser/utils/scheme_logger.cc

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

I want to know about https://bugs.chromium.org/p/chromium/issues/detail?id=1467134 have any good ideas?

FYI. Intent to Ship is here (by @gjc).
https://groups.google.com/a/chromium.org/g/blink-dev/c/wYuPrIQzDTA/m/uoL4bXR2BgAJ

[Empty comment from Monorail migration]

Currently, there are multiple issues reports related to non-special URL:

https://crbug.com/1416001
https://crbug.com/1416006
https://crbug.com/1416020
https://crbug.com/1416022
https://crbug.com/869291

It seems some of these issues overlap.

To streamline issue management and tracking, I would like to merge these reports into a single issue:

https://crbug.com/1416006

If I find it's more appropriate to create a sub issue for specific aspects, I'll do so.
For now, consolidating these reports into one issue will help us track and address potential root causes more effectively.

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

[Empty comment from Monorail migration]

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/67985d6809954707cdcdf17698f36254c1f45d9b

commit 67985d6809954707cdcdf17698f36254c1f45d9b
Author: Hayato Ito <hayato@chromium.org>
Date: Thu Oct 26 01:11:29 2023

Refactor url_parse_unittest.cc

This CL is pure refacotring, to prepare for https://crbug.com/1416006.
No behavior is changed.

Suppress clang-tidy warnings by fixing the followings:

- [modernize-use-nullptr]: Use nullptr
- [modernize-loop-convert]: Use range-based for loop instead

Bug: 1416006

Change-Id: I13f7297396b6800508a65f48fd72b4890964e428
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4975082
Auto-Submit: Hayato Ito <hayato@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Reviewed-by: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1215252}

[modify] https://crrev.com/67985d6809954707cdcdf17698f36254c1f45d9b/url/url_parse_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/14e638ecc947e45be51df41bdd1b624bba12eed3

commit 14e638ecc947e45be51df41bdd1b624bba12eed3
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Oct 30 02:22:58 2023

Refactor url_util_unittest.cc

This CL is pure refacotring, to prepare for https://crbug.com/1416006.
No behavior is changed.

Suppress clang-tidy warnings by fixing the followings:

- [modernize-use-nullptr]: Use nullptr
- [modernize-loop-convert]: Use range-based for loop instead

Bug: 1416006
Change-Id: If44052025fbbb2cd9b4701bd8b50775f85ab19ba
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4975287
Reviewed-by: Adam Rice <ricea@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1216750}

[modify] https://crrev.com/14e638ecc947e45be51df41bdd1b624bba12eed3/url/url_util_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/d311bb222cf8924118c336c8de95d4e7c44e6893

commit d311bb222cf8924118c336c8de95d4e7c44e6893
Author: Hayato Ito <hayato@chromium.org>
Date: Fri Nov 10 01:35:55 2023

Add a feature flag for standard-compliant non-special scheme URL parsing

This CL is separate from the main CL (https://crrev.com/c/4929491).
As a first step, this CL adds a URL feature flag and a virtual test with
it.

Bug: 1416006
Change-Id: I12533902bffff2e74bd48b98833c1dd13aeddf5d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5015730
Commit-Queue: Hayato Ito <hayato@chromium.org>
Reviewed-by: Koji Ishii <kojii@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1222683}

[add] https://crrev.com/d311bb222cf8924118c336c8de95d4e7c44e6893/third_party/blink/web_tests/virtual/url-non-special/OWNERS
[add] https://crrev.com/d311bb222cf8924118c336c8de95d4e7c44e6893/third_party/blink/web_tests/virtual/url-non-special/README.md
[modify] https://crrev.com/d311bb222cf8924118c336c8de95d4e7c44e6893/third_party/blink/web_tests/VirtualTestSuites
[modify] https://crrev.com/d311bb222cf8924118c336c8de95d4e7c44e6893/url/url_features.h
[modify] https://crrev.com/d311bb222cf8924118c336c8de95d4e7c44e6893/url/url_features.cc

I'm starting to write an explainer:

https://docs.google.com/document/d/1LjxHl32fE4tCKugrK_PIso7mfXQVEeoD1wSnX2y0ZU8/edit?usp=sharing&resourcekey=0-d1gP4X2sG7GPl9mlTeptIA

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/2579d83bbdf1b6893506568fce4916ead75715c4

commit 2579d83bbdf1b6893506568fce4916ead75715c4
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Nov 29 07:57:40 2023

Rename IsURLSlash to IsSlashOrBackslash

This CL is separate from https://crrev.com/c/5028539.
This is a pure renaming, no behavior change.

Bug: 1416006
Change-Id: Id861375a9b7344f3775246fbf9991752ec28c5c5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5068436
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1230486}

[modify] https://crrev.com/2579d83bbdf1b6893506568fce4916ead75715c4/url/url_file.h
[modify] https://crrev.com/2579d83bbdf1b6893506568fce4916ead75715c4/url/url_canon_path.cc
[modify] https://crrev.com/2579d83bbdf1b6893506568fce4916ead75715c4/url/url_parse_internal.h
[modify] https://crrev.com/2579d83bbdf1b6893506568fce4916ead75715c4/url/third_party/mozilla/url_parse.cc
[modify] https://crrev.com/2579d83bbdf1b6893506568fce4916ead75715c4/url/url_canon_relative.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/b37956bf218bc0780bb7264c21bb41b4d7e07353

commit b37956bf218bc0780bb7264c21bb41b4d7e07353
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Dec 04 05:01:52 2023

Add non-special scheme URLs parser

This CL is separate from the main CL (https://crrev.com/c/4929491).

This CL adds a URL parser for non-special URLs.

The newly added parser functions are currently unused and will be used
in following CLs. Thus, guarding these functions by the flag is
unnecessary at this point.

As described in `url/README.md`, URL library has two layers:

1. Parsing
2. Canonicalization

This CL updates "Parsing" layer for non-special URLs.

Canonicalization of non-special URLs is a more complex task. It will be
addressed in following CLs.

Tests:

This CL adds minimal unit tests for the ParseNonSpecialURL() function.
These tests should be sufficient to verify the basic behavior of the
parser.

The WPT URL tests have a good coverage for non-special URLs and will be
used in following CLs, so I don't add much unit tests in this CL to
avoid duplication.

Bug: 1416006
Change-Id: Ibe5450ccdd4fc6d5c2386b1e838e9e254ce90308
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5028539
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Timothy Gu <timothygu@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232562}

[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/url_parse_unittest.cc
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/components/url_formatter/url_fixer.cc
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/url_parse_internal.h
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/README.md
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/third_party/mozilla/url_parse.h
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/third_party/mozilla/url_parse.cc
[modify] https://crrev.com/b37956bf218bc0780bb7264c21bb41b4d7e07353/url/url_canon_relative.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/203c30248db0dc01c52429ea2664a139396f2203

commit 203c30248db0dc01c52429ea2664a139396f2203
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Dec 11 04:58:57 2023

[URL] Canonicalize hosts in non special URLs

This CL, separate from the main CL (https://crrev.com/c/4929491), adds
several functions to canonicalize hosts in non-special URLs,
implementing the following URL Standard sections:

- https://url.spec.whatwg.org/#host-parsing
- https://url.spec.whatwg.org/#concept-opaque-host-parser

The newly added functions are currently unused and will be used
in following CLs. Thus, guarding these functions by the flag is
unnecessary at this point.

Tests:

This CL adds minimal unit tests for CanonicalizeNonSpecialHost(...)
functions. These tests should be sufficient to verify the basic behavior
of the canonicalization functions.

The WPT URL tests have a good coverage for non-special URLs and will be
used in following CLs.

Bug: 1416006

Change-Id: I762053754bcf3417e277e660a9d83833573ddb2c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5092012
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1235585}

[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon.h
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon_host.cc
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon_pathurl.cc
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon_unittest.cc
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/third_party/mozilla/url_parse.h
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon_internal.h
[modify] https://crrev.com/203c30248db0dc01c52429ea2664a139396f2203/url/url_canon_ip.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/727115f21900e7b5851854e9233c02d59534fba9

commit 727115f21900e7b5851854e9233c02d59534fba9
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Dec 11 06:41:58 2023

[URL] Support non-special URLs in CanonicalizePath

This CL, separate from the main CL (https://crrev.com/c/4929491),
updates the canonicalize functions for URL paths to consider non-special
URLs, implementing the following URL Standard sections:

- https://url.spec.whatwg.org/#path-start-state
- https://url.spec.whatwg.org/#path-state

The URL Standard is a bit difficult to read, so here is the summary of
the changes for non-special URLs:

- Backslashes (\) are no longer treated as path separators.
- The canonical path is empty instead of a single slash (/) if no path
  component is present.

Bug: 1416006
Change-Id: I18303f021cc27eea6f2b33ba3dc575a9f512c26c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5095842
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1235601}

[modify] https://crrev.com/727115f21900e7b5851854e9233c02d59534fba9/url/url_canon.h
[modify] https://crrev.com/727115f21900e7b5851854e9233c02d59534fba9/url/url_canon_path.cc
[modify] https://crrev.com/727115f21900e7b5851854e9233c02d59534fba9/url/url_canon_unittest.cc
[modify] https://crrev.com/727115f21900e7b5851854e9233c02d59534fba9/url/url_canon_internal.h
[modify] https://crrev.com/727115f21900e7b5851854e9233c02d59534fba9/url/url_canon_relative.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3

commit 735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Dec 13 02:49:38 2023

[URL] Canonicalize non-special URLs

Add CanonicalizeNonSpecialURL functions, putting the followings into
together:

- https://crrev.com/c/5028539 (Parser for non-special URLs)
- https://crrev.com/c/5092012 (Canonicalize non-special hosts)
- https://crrev.com/c/5095842 (Canonicalize non-special paths)

This CL is separate from the main CL (https://crrev.com/c/4929491).

Bug: 1416006
Change-Id: If7125ff5f052e465dcaa5876061690697899872b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5114580
Reviewed-by: Timothy Gu <timothygu@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1236752}

[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_canon.h
[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_canon_fileurl.cc
[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_parse_unittest.cc
[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/BUILD.gn
[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_canon_stdurl.cc
[add] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_canon_non_special_url.cc
[modify] https://crrev.com/735a5cbcef5c4d88e84b2258ad6cfdf05300e2a3/url/url_canon_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/08f24b9178f337a0ce50c143daab639cd85e3169

commit 08f24b9178f337a0ce50c143daab639cd85e3169
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Dec 13 10:01:41 2023

[URL] Remove unused parameters

Remove unused parameters and template parameters from:

- url::CanonicalizeStandardURL
- url::CanonicalizeFileSystemURL
- DoCanonicalizeFileSystemURL

Bug: 1416006
Change-Id: I40f29d53c947aae6c09b3809a2378ea19c212a0f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5114701
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Adam Rice <ricea@chromium.org>
Reviewed-by: Viktor Semeniuk <vsemeniuk@google.com>
Commit-Queue: Viktor Semeniuk <vsemeniuk@google.com>
Auto-Submit: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1236868}

[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_util.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_canon.h
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_parse_perftest.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/components/password_manager/core/browser/affiliation/affiliation_utils.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/net/dns/public/dns_over_https_server_config.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_canon_stdurl.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_canon_unittest.cc
[modify] https://crrev.com/08f24b9178f337a0ce50c143daab639cd85e3169/url/url_canon_filesystemurl.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/667ad268cecc08965eeb12ea56f79f3fa3a24d0d

commit 667ad268cecc08965eeb12ea56f79f3fa3a24d0d
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Jan 24 08:45:39 2024

Add a comment to mention the EarlyAccess allow list

See https://crbug.com/1520386 for more details.

Bug: 1520386,1416006
Change-Id: Ia50a361faed473500331dcabb15d8910dbe31e55
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5232740
Reviewed-by: Adam Rice <ricea@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1251298}

[modify] https://crrev.com/667ad268cecc08965eeb12ea56f79f3fa3a24d0d/url/url_features.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/f3ae9cad5b60b2ef99d117d298f916f21c29657c

commit f3ae9cad5b60b2ef99d117d298f916f21c29657c
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Jan 29 07:20:01 2024

[URL] Carry over parsed.host correctly in canonicalizing url's host

The current code incorrectly assigns an invalid Component ({0, -1}) to
`out_host` during URL canonicalization, even when the input host is
empty (e.g. {6, 0}). Fix that because the difference matters in
non-special URLs.

This is a blocking issue for https://crrev.com/c/5222760.

Bug: 1416006
Change-Id: I6894aa3af6819c70e5cc8feb8ba64ef3c52e00f3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5231566
Reviewed-by: Timothy Gu <timothygu@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1253173}

[modify] https://crrev.com/f3ae9cad5b60b2ef99d117d298f916f21c29657c/url/url_canon_host.cc
[modify] https://crrev.com/f3ae9cad5b60b2ef99d117d298f916f21c29657c/url/url_canon_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/0f402c037c5a86d11eafe654dd454a6e6cac9a5c

commit 0f402c037c5a86d11eafe654dd454a6e6cac9a5c
Author: Hayato Ito <hayato@chromium.org>
Date: Mon Jan 29 07:53:13 2024

[URL] Support non-special URLs in Resolve()

Implement correct support for non-special URLs in gurl::Resolve() and
other functions, which are used on resolving a relative URL on a given
base URL.

Example:

Prior to this change:

> new URL("path", "git://host/") => throws an invalid URL error

After this change:

> new URL("path", "git://host/") => "git://host/path""

The feature is currently guarded by the flag.

This CL adds several parameterized unit tests to test flag-dependent
behaviors. Currently flag-independent behaviors are not tested with
the flag. I will revisit this later.

Though this CL introduces new expectation files for virtual tests,
there is no regression and the total number of test failures has
decreased. I use rebaseline cl tool and accept the result as is. It
seems there are several expectation files across different platforms,
accumulated over time.


Bug: 1416006
Change-Id: Ib52e9977867befa1483f7821c469a4dd8f427627
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5222760
Reviewed-by: Timothy Gu <timothygu@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1253180}

[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/url/url_util.cc
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/flag-specific/disable-site-isolation-trials/webaudio/internals/audioworkletprocessor-gc.https-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/url/gurl_unittest.cc
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/flag-specific/disable-site-isolation-trials/virtual/url-non-special/external/wpt/url/a-element-origin-xhtml-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/url/url_canon_relative.cc
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any.worker-expected.txt
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/url/url_util_unittest.cc
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any-expected.txt
[modify] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/url/url_canon_unittest.cc
[add] https://crrev.com/0f402c037c5a86d11eafe654dd454a6e6cac9a5c/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-xhtml-expected.txt

CL from https://crbug.com/chromium/1416006#c37 (https://crrev.com/c/5222760) has caused consisten failures on linux bots.

--- /b/s/w/io7pu4bthf/layout-test-results/webaudio/internals/audioworkletprocessor-gc.https-expected.txt
+++ /b/s/w/io7pu4bthf/layout-test-results/webaudio/internals/audioworkletprocessor-gc.https-actual.txt
@@ -1,5 +1,3 @@
 This is a testharness.js-based test.
-[FAIL] Test GC of AudioWorkletProcessor
-  assert_equals: the initial handler count should be zero. expected 0 but got 5
 Harness: the test ran to completion.
 

I will go ahead and revert the CL.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/54148048e463a6945fadba1f885856c388e64e78

commit 54148048e463a6945fadba1f885856c388e64e78
Author: Alexander Hendrich <hendrich@chromium.org>
Date: Mon Jan 29 10:48:19 2024

Revert "[URL] Support non-special URLs in Resolve()"

This reverts commit 0f402c037c5a86d11eafe654dd454a6e6cac9a5c.

Reason for revert: fails on linux bot (https://crbug.com/1416006#c38)

Original change's description:

Bug: 1416006
Change-Id: I839ec6d567bbe36351c8c530526113c1e94c7f07
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5244746
Auto-Submit: Alexander Hendrich <hendrich@chromium.org>
Owners-Override: Alexander Hendrich <hendrich@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Alexander Hendrich <hendrich@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1253214}

[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/url/url_util.cc
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/flag-specific/disable-site-isolation-trials/webaudio/internals/audioworkletprocessor-gc.https-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/url/gurl_unittest.cc
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/flag-specific/disable-site-isolation-trials/virtual/url-non-special/external/wpt/url/a-element-origin-xhtml-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/url/url_canon_relative.cc
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any.worker-expected.txt
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/url/url_util_unittest.cc
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any-expected.txt
[modify] https://crrev.com/54148048e463a6945fadba1f885856c388e64e78/url/url_canon_unittest.cc
[delete] https://crrev.com/ca8f62c7f553b35a1b03701945532f4a800610f9/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-xhtml-expected.txt

The CL also caused other failures on two differing linux dbg bots:
https://ci.chromium.org/ui/p/chromium/builders/ci/linux-chromeos-dbg/35470/test-results?sortby=&groupby=
https://ci.chromium.org/ui/p/chromium/builders/ci/Linux%20Tests%20(dbg)(1)/118400/test-results?sortby=&groupby=

Failure reason:
gurl.cc(120): Check failed: test_url.parsed_.host == parsed_.host ({0, -1} vs. {7, 0})

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/b505aca89dc9b2860a0bb480c19d571faacb67be

commit b505aca89dc9b2860a0bb480c19d571faacb67be
Author: Joshua Hood <jdh@chromium.org>
Date: Mon Jan 29 13:39:32 2024

Revert "[URL] Carry over parsed.host correctly in canonicalizing url's host"

This reverts commit f3ae9cad5b60b2ef99d117d298f916f21c29657c.

Reason for revert: https://crbug.com/1522816

Original change's description:

Bug: 1416006
Change-Id: Ic53dda4105bd01db1558e26476d3233ec37f280b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5245065
Owners-Override: Joshua Hood <jdh@chromium.org>
Auto-Submit: Joshua Hood <jdh@chromium.org>
Commit-Queue: Joshua Hood <jdh@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#1253259}

[modify] https://crrev.com/b505aca89dc9b2860a0bb480c19d571faacb67be/url/url_canon_host.cc
[modify] https://crrev.com/b505aca89dc9b2860a0bb480c19d571faacb67be/url/url_canon_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/bbdc90b76304d8fa34b9b65e9b942e88352519e3

commit bbdc90b76304d8fa34b9b65e9b942e88352519e3
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Jan 31 02:57:28 2024

Reland "[URL] Carry over parsed.host correctly in canonicalizing url's host"

This is a reland of commit f3ae9cad5b60b2ef99d117d298f916f21c29657c

The diff to the reveted CL is https://crrev.com/c/5247368/1..3

The reverted CL caused a DCHECK() error [1] inside the NDEBUG block of
the GURL sanitizer check. It is likely that the CQ try bots do not
include any debug bots by default.

I ran linux_chromium_dbg_ng in CQ try too this time.

Although I added a comment why we need to handle Component(0, 0)
specially, it's a bit complicated topic. I will revisit this when I
update url::Origin and other code to support empty hosts of
non-special URLs in a more robust manner.

- [1] https://source.chromium.org/chromium/chromium/src/+/main:url/gurl.cc;l=120


Original change's description:

Bug: 1416006
Change-Id: I097f39f2cad8a20d13e242594734bf8f9428b853
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5247368
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1254331}

[modify] https://crrev.com/bbdc90b76304d8fa34b9b65e9b942e88352519e3/url/url_canon_host.cc
[modify] https://crrev.com/bbdc90b76304d8fa34b9b65e9b942e88352519e3/url/url_canon_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src/+/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0

commit 1087d8e07c2948dbcd195cd7c700037b4b5bc3b0
Author: Hayato Ito <hayato@chromium.org>
Date: Wed Jan 31 05:48:13 2024

Reland "[URL] Support non-special URLs in Resolve()"

This is a reland of commit 0f402c037c5a86d11eafe654dd454a6e6cac9a5c

The diff to the reveted CL is https://crrev.com/c/5246969/1..3

I removed expectation files under
flag-specific/disable-site-isolation-trials that were introduced by a
previous 'rebaseline cl'.

While the tool aims to update expectations accurately, it seems
manual verification is still needed.

Original change's description:

Bug: 1416006
Change-Id: I60b2a26fb796a19a20034055abba3214e50ebaee
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5246969
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Hayato Ito <hayato@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1254363}

[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/url/url_util.cc
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/url/gurl_unittest.cc
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/url/url_canon_relative.cc
[add] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any.worker-expected.txt
[add] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/url/url_util_unittest.cc
[add] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-origin.any-expected.txt
[modify] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/url/url_canon_unittest.cc
[add] https://crrev.com/1087d8e07c2948dbcd195cd7c700037b4b5bc3b0/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-origin-xhtml-expected.txt

This issue was migrated from crbug.com/chromium/1416006?no_tracker_redirect=1

[Monorail blocking: crbug.com/chromium/1058053, crbug.com/chromium/1412660, crbug.com/chromium/1416001, crbug.com/chromium/1416018, crbug.com/chromium/1416022]
[Monorail mergedwith: crbug.com/chromium/1416001, crbug.com/chromium/1416014, crbug.com/chromium/1416020, crbug.com/chromium/1416022, crbug.com/chromium/1464448, crbug.com/chromium/1467134, crbug.com/chromium/1487830, crbug.com/chromium/869291]
[Monorail components added to Component Tags custom field.]

Project: chromium/src
Branch: main

commit be97d681ef495b558cf1bc64ba296d1a7b6d42d2
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Feb 06 09:40:57 2024

    [URL] Support non-special URLs in ReplaceComponents()
   
    Implement correct support for non-special URLs in
    gurl::ReplaceComponents().
   
    Example:
   
    Prior to this change:
   
    > const url = new URL("git://host/path")
    > url.host = "hello";
    > url.pathname = "world";
    > url.href
    "git://host/path"
   
    After this change:
   
    > const url = new URL("git://host/path")
    > url.host = "hello";
    > url.pathname = "world";
    > url.href
    "git://hello/world"
   
    This CL also adds ParseNonSpecialURLInternal(...) functions with
    `trim_path_end` parameter, which is required to support the following
    tricky case, preserving a trailing whitespace in paths when removing a
    hash.
   
    > const url = new URL("git://host/path #a")
    > url.hash = "";
    > url.href
    "git://host/path%20"
   
    Note: The existing ParsePathURL(...) has a similar `trim_path_end` parameter.
   
    Bug: 1416006
    Change-Id: Ibc7aaed7a9b4a1ab608eb1d90e961f1cc8921c0c
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5240153
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Reviewed-by: Timothy Gu <timothygu@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1256669}

M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_include=javascript-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-stripping.any-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-stripping.any.worker-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_include=javascript-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_include=javascript-expected.txt
M       url/gurl_unittest.cc
M       url/third_party/mozilla/url_parse.cc
M       url/url_canon.h
M       url/url_canon_internal.cc
M       url/url_canon_non_special_url.cc
M       url/url_canon_unittest.cc
M       url/url_parse_internal.h
M       url/url_util.cc
M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5240153

Project: chromium/src
Branch: main

commit f78f2dae48473a10e962dab92a5b725b75a74001
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Feb 06 22:56:40 2024

    [URL] Removes a test case which was moved to flag-dependent tests
   
    The test case was already migrated to flag-dependent tests in a
    previous CL, https://crrev.com/c/5246969.
   
    I forgot to remove the test case from the original location.
   
    Bug: 1416006
    Change-Id: I5ab40095d6a79e69040ccc2d9a2d208d7cb243b1
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5271866
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Commit-Queue: Daniel Cheng <dcheng@chromium.org>
    Auto-Submit: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1257048}

M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5271866

Project: chromium/src
Branch: main

commit a2bacc3a698bc74d61ff47ca4039dbf245d6953a
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Feb 09 07:02:51 2024

    [URL] Migrate some tests in gurl_unittest to flag-dependent tests
   
    Bug: 40063064
    Change-Id: Iff261b92d88cc22dd1666d0c5be246302c482fce
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5273330
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Timothy Gu <timothygu@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1258379}

M       url/gurl.h
M       url/gurl_unittest.cc

https://chromium-review.googlesource.com/5273330

Project: chromium/src
Branch: main

commit a5bb75e1fdd2f34e49038998af3844f3f9f7b9b5
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Feb 09 07:19:04 2024

    [URL] Disallow setting user/password/port for URLs with empty hosts
   
    This CL aligns with the URL Standard by preventing the setting of
    user, password, or port for URLs that don't permit them, specifically
    when the host is null or empty.
   
    This is a follow-up of the previous review:
    https://crrev.com/c/5240153/15/third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_exclude%3D(file_javascript_mailto)-expected.txt#20
   
    Prior to this change:
   
    > const url = "git:///";
    > url.username = "x";
    > url.href
    "git://x@/"
   
    After this change:
   
    > const url = "git:///";
    > url.username = "x";
    > url.href
    "git:///"
   
    The URL Standard:
    - https://url.spec.whatwg.org/#cannot-have-a-username-password-port
   
    Username/password/port setters:
    - https://url.spec.whatwg.org/#dom-url-username
    - https://url.spec.whatwg.org/#dom-url-password
    - https://url.spec.whatwg.org/#dom-url-port
   
    Since non-special URLs can now have a null or empty host, we should
    consider a case where users set username/password/port for valid URLs
    that have a null or empty host.
   
    First, I attempted to update url::ReplaceComponents() so we can return
    early, but GURL::ReplaceComponents() and KURL::ReplaceComponents() are
    using url::ReplaceComponents() differently. I couldn't find a good
    way.
   
    Next, I considered updating both GURL::ReplaceComponents() and
    KURL::ReplaceComponents() so they return early without calling
    url::ReplaceComponents(), but redundancy was a concern.
   
    After that, I realized that the best place to address this is the
    canonicalization function for non-special URLs,
    DoCanonicalizeNonSpecialURL(), as canonicalization occurs after URL
    replacements, anyway. Implementing proper canonicalization should
    effectively resolve the issue.
   
    Bug: 40063064
    Change-Id: Ibeb7f2dbaad196d83d26c8fb87bf47496bc989e5
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5276359
    Reviewed-by: Timothy Gu <timothygu@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1258381}

M       third_party/blink/renderer/core/url/dom_url_utils.cc
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_exclude=(file_javascript_mailto)-expected.txt
M       url/gurl_unittest.cc
M       url/url_canon_non_special_url.cc
M       url/url_parse_unittest.cc

https://chromium-review.googlesource.com/5276359

Project: chromium/src
Branch: main

commit 826a8158bd76bbe6ced7ef25dad6ee74deac8948
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Feb 19 07:55:06 2024

    [URL] Migrate net/base/url_util_unittest to flag-dependent tests
   
    Bug: 40063064
    Change-Id: I561037692c18d97f509de765f87243897be678b0
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5298419
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1262269}

M       net/base/url_util_unittest.cc

https://chromium-review.googlesource.com/5298419

Project: chromium/src
Branch: main

commit 826a8158bd76bbe6ced7ef25dad6ee74deac8948
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Feb 19 07:55:06 2024

    [URL] Migrate net/base/url_util_unittest to flag-dependent tests
   
    Bug: 40063064
    Change-Id: I561037692c18d97f509de765f87243897be678b0
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5298419
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1262269}

M       net/base/url_util_unittest.cc

https://chromium-review.googlesource.com/5298419

Project: chromium/src
Branch: main

commit 2eb072ba0f723dfff81b923249ef9e7bbb1dae57
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Feb 19 09:01:18 2024

    Add non-special URL test cases to scheme_host_port_unittest
   
    Add several test cases to increase test coverage.
   
    Bug: 40063064
    Change-Id: Ifefb968a4fc9c7cfee7808ec90e2e5cfd47c8acf
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5302098
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1262279}

M       url/scheme_host_port_unittest.cc

https://chromium-review.googlesource.com/5302098

Project: chromium/src
Branch: main

commit 2eb072ba0f723dfff81b923249ef9e7bbb1dae57
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Feb 19 09:01:18 2024

    Add non-special URL test cases to scheme_host_port_unittest
   
    Add several test cases to increase test coverage.
   
    Bug: 40063064
    Change-Id: Ifefb968a4fc9c7cfee7808ec90e2e5cfd47c8acf
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5302098
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1262279}

M       url/scheme_host_port_unittest.cc

https://chromium-review.googlesource.com/5302098

Project: chromium/src
Branch: main

commit 170aec2a2d5cd73e07bfb0aae09374ec8bc3cab5
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Feb 21 04:09:59 2024

    [URL] Migrate url/url_util_unittest to flag-dependent tests
   
    URLUtilTest::TestNoRefComponent test was originally written before
    full support for non-special URLs became available. We need a
    flag-dependent test here because the test uses an internal parse
    function.
   
    The test case corresponds to the following user scenario:
   
    > const url = new URL("any#body", "mailto://to/");
    > assertEquals(url.href, "mailto://to/any#body");
   
    Bug: 40063064
    Change-Id: Icca4260f4646368e19412d6e9009b7098cfd0d37
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5298936
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1263130}

M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5298936

Project: chromium/src
Branch: main

commit d3c90ed5528623aa64aab9a42c4039c08453cff6
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Mar 05 07:31:34 2024

    [URL] Enable setting a host, port, or pathname on non-special URLs
   
    This CL, a follow-up to https://crrev.com/c/5276359, allows Web facing
    URL APIs to set a host, port, or pathname, on non-special URLs.
   
    This functionality was previously available for internal URL structures
    through prior CLs, but not exposed to JS APIs.
   
    The CL updates KURL::IsHierarchical() to recognize non-special URLs as
    hierarchical and introduces KURL::CanRemoveHost() to handle an empty
    host correctly.
   
    Maintaining the current behavior with the feature flag disabled while
    introducing these changes might slightly impact code readability,
    unfortunately.
   
    Examples:
   
    Before:
   
    > const url = new URL("git://h/")
    > url.host = "";
    > url.href
    "git://h/"
   
    After:
   
    > const url = new URL("git://h/")
    > url.host = "";
    > url.href
    "git:///"
   
    Before:
   
    > const url = new URL("git://h/");
    > url.port = 80;
    > url.href
    "git://h/"
   
    After:
   
    > const url = new URL("git://h/")
    > url.port = 80;
    > url.href
    "git://h:80/"
   
    Note: Setting an empty host is disallowed when a port exists, as
    follows:
   
    > const url = new URL("git://h:80/")
    > url.host = "";
    > url.href
    "git://h:80/"
   
    Tests:
   
    The WPT url tests has a good coverage of these test cases. Many
    previously failing tests are now passing.
   
    Some tests involving `pathname` are still failing. These failures occur
    because `pathname` processing should differ between special and
    non-special URLs. That will be addressed in a separate follow-up CL.
   
   
    Bug: 40063064
    Change-Id: I0d71921edfa5f13ae60b8ce7ea8117d5fdeef7de
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5318947
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Timothy Gu <timothygu@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1268290}

M       third_party/blink/renderer/core/url/dom_url_utils.cc
M       third_party/blink/renderer/platform/weborigin/kurl.cc
M       third_party/blink/renderer/platform/weborigin/kurl.h
M       third_party/blink/renderer/platform/weborigin/kurl_test.cc
M       third_party/blink/renderer/platform/weborigin/security_origin.cc
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_exclude=(file_javascript_mailto)-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_include=file-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_include=javascript-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-stripping.any-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-stripping.any.worker-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_exclude=(file_javascript_mailto)-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_include=file-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_include=javascript-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_exclude=(file_javascript_mailto)-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_include=file-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_include=javascript-expected.txt

https://chromium-review.googlesource.com/5318947

Project: chromium/src
Branch: main

commit cddb0878c7ed93f79684b845751f35eeb4794528
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Mar 13 01:02:24 2024

    [URL] Update Origin and SecurityOrigin to consider non-special URLs
   
    Update url::Origin and blink::SecurityOrigin to consider the correct
    support of non-special scheme URLs (http://crbug.com/40063064), by which
    host and port information are now available even for non-special URLs.
   
    Background:
   
    The Web Standard requires non-special URLs to have an opaque origin.
    Chromium generally follows this standard. Thus, regarding Origin, the
    correct support of non-special URLs should have minimal impact on web
    compatibility because host and port parts are unimportant to opaque
    origins.
   
    However, due to historical reasons, Chromium has implemented several
    workarounds for non-special scheme URLs, such as `url::SchemeRegistry`
    and AndroidWebViewHack.
   
    Example:
   
    url::SchemeRegistry allows non-special URLs to be treated as if they
    have non-opaque origins, as if they were standard URLs in parsing them,
    and so on.
   
    Suppose that we register "local-but-nonstandard:" scheme as a Local
    scheme using `url::AddLocalScheme("local-but-nonstandard")`.
   
    Then, the origin of "local-but-nonstandard://examplecom:80" URL becomes
    non-opaque.
   
    However, given that "local-but-nonstandard:" scheme is not registered as
    a standard scheme, the origin's internal `SchemeHostPort` tuple will be:
   
    ("local-but-nonstandard", "", 0)
   
    instead of
   
    ("local-but-nonstandard", "example.com", 80)
   
    because of incomplete parsing of non-special URLs so far, where host and
    port information are not captured.
   
    Changes:
   
    Now, we have a correct non-special URL scheme parsing. The origin's
    internal SchemeHostPort tuple can be
   
    ("local-but-nonstandard", "example.com", 80)
   
    because the new parser parses non-special URLs correctly.
   
    This CL updates url::Origin so that it considers non-special URLs
    correctly, aligning with the correct non-special URL support.
   
    Impacts:
   
    While the overall web compatibility impact is expected to be minimal,
    there might be some unintended consequences in Chromium specific
    scenarios, such as:
   
    - "local-but-nonstandard:" scheme behavior: Code relying on this
      scheme's previous non-opaque behavior might be affected due to changes
      in origin handling.
   
    - AndroidWebViewHack: Potential compatibility issues with the Android
      WebView component, which is extremely high.
   
    Ideally, we should update the current behavior everywhere so that we can
    consider host and port parts, instead of ignoring them, however, this
    would bring an extremely high risk of breakage for Android WebView.
    There is basically no way to change how Android WebView behaves.
   
    Therefore, this CL updates the behaviors of "local-but-nonstandard"
    cases, but preserves the current Android WebView behavior by discarding
    host and port information intentionally. This might be a bad outcome.
   
    This is not a final decision. I'll re-evaluate once I have more
    information.
   
   
    The summary of what's changed and what's not changed:
   
    For the two URLs, A) "non-standard://a.com:80" and B)
    "non-standard://b.com:80", and with AndroidWebViewHack enabled:
   
    ```
    |                  | Same origin? | Serialization of A's origin |
    |------------------+--------------+-----------------------------|
    | Current behavior | true         | "non-standard://"           |
    | After this CL    | true         | "non-standard://"           |
    ```
   
   
    For the two URLs, A) "local-but-nonstandard://a.com:80" and B)
    "local-but-nonstandard://b.com:80":
   
    ```
    |                  | Same origin? | Serialization of A's origin        |
    |------------------+--------------+------------------------------------|
    | Current behavior | true         | "local-but-nonstandard://"         |
    | After this CL    | false        | "local-but-nonstandard://a.com:80" |
    ```
   
    See url/origin_unittest.cc and url/origin_abstract_tests.h in this CL
    for details.
   
    For historical context on nonstandard scheme URLs and Andvoid WebView
    behavior, see the followings:
   
    - https://crrev.com/c/1208811
    - https://crrev.com/c/1338660
   
    WPT url Tests:
   
    There should be no impacts on WPT url tests.
   
    Next steps:
   
    Following CLs will address potential issues in tests and code currently
    relying on the old workarounds. These subsequent changes will help us
    assess the overall impact and ensure a smooth transition to the new
    behavior.
   
    Bug: 40063064
    Change-Id: I53ac86503041b869a002a1ea02b18e0002b4b11d
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5309015
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1271918}

M       android_webview/java/src/org/chromium/android_webview/common/ProductionSupportedFlagList.java
M       third_party/blink/renderer/platform/weborigin/security_origin.cc
M       third_party/blink/renderer/platform/weborigin/security_origin.h
M       url/origin.cc
M       url/origin_abstract_tests.h
M       url/origin_unittest.cc
M       url/scheme_host_port.cc
M       url/scheme_host_port.h
M       url/url_util.cc
M       url/url_util.h

https://chromium-review.googlesource.com/5309015

Project: chromium/src
Branch: main

commit 234328a0b4a4f5067673a0fcb4cdba7a78dec8c2
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Mar 15 02:15:55 2024

    [URL] Support resolving an opaque path with non-special base URLs
   
    Fix the wrong result in resolving an opaque path with non-special base
    URLs.
   
    Before:
   
    > const url = new URL("git:opaque", "git://host/a") ;
    > url.href
    => "git://host/opaque"
   
    After:
   
    > const url = new URL("git:opaque", "git://host/a") ;
    > url.href
    => "git:opaque"
   
   
    URL Standard says "if url is special" in 2.6 of scheme-state:
   
    https://url.spec.whatwg.org/#scheme-state
   
    > 2.6. Otherwise, if url is special, base is non-null, and base’s scheme
    > is url’s scheme:
   
    However, our implementation didn't implement this condition check.
   
    While the WPT URL tests have a good coverage for these scenarios, I also
    added unit tests for additional assurance.
   
    Bug: 40063064
    Change-Id: Ic9f9b4c3b55bd2cac0c8993638e55fc8d9cf8180
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5364258
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1273193}

M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
M       url/gurl_unittest.cc
M       url/url_canon_relative.cc

https://chromium-review.googlesource.com/5364258

Project: chromium/src
Branch: main

commit ec4bb766bb9f04e153aa0f2ad6909f9c298fe350
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Mar 15 04:33:09 2024

    [URL] Support empty hosts in relative URLs for non-special base URLs
   
    Update Chromium's handling of relative URLs by allowing them to have
    empty hosts (like "///path") when used with non-special base URLs.
   
    Before:
   
    > const url = new URL("///path", "git://host/a") ;
    => throws an Invalid URL error.
   
    After:
   
    > const url = new URL("///path", "git://host/a") ;
    => `url' is valid. url.href is "git:///path".
   
    This change basically implements "host state" in URL Standard:
    https://url.spec.whatwg.org/#host-state
   
    > 2.1. If url is special and buffer is the empty string, host-missing
    > validation error, return failure.
   
    Prior CLs already implemented this, but was not implemented correctly in resolving relative URLs for non-special base URLs.
   
    While the WPT URL tests have a good coverage for these scenarios, I also
    added unit tests for additional assurance.
   
    Bug: 40063064
    Change-Id: I46786955d5db27d29864e80a259ce4076500cb33
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5364453
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1273227}

M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
M       url/gurl_unittest.cc
M       url/third_party/mozilla/url_parse.cc
M       url/url_canon_relative.cc
M       url/url_parse_internal.h

https://chromium-review.googlesource.com/5364453

Project: chromium/src
Branch: main

commit 6957f7755b6e190a42676a199e856c7dc09c737f
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Mar 18 03:00:46 2024

    [URL] Ensure correct handling of path-only non-special URLs
   
    Prevent a path-only non-special URL, "non-special:/.//path", from
    ending up as "non-special://path", instead of "non-special:/.//path".
   
    Note that a leading "/." should be *prepended* before the path so the
    path part, "//path", shouldn't be considered as a URL's host part.
   
    Previously, these URLs might incorrectly lose the leading "./" part, or
    forget to add "./".
   
    This implements the URL Standard: 4.5 serializing:
   
    https://url.spec.whatwg.org/#url-serializing
   
    > 3. If url’s host is null, url does not have an opaque path, url’s
    > path’s size is greater than 1, and url’s path[0] is the empty string,
    > then append U+002F (/) followed by U+002E (.) to output.
   
    Examples:
   
    > const url = new URL("git:/.//a");
    > url.href
    => should be "git:/.//a", instead of "git://a".
   
    > const url = new URL("git:/");
    > url.pathname = "//a";
    > url.href
    => should be "git:/.//a", instead of "git://a".
   
    > const url = new URL("", "git:/.//a");
    > url.href
    => should be "git:/.//a", instead of "git://a".
   
    > const url = new URL("..//b", "git:/.//a");
    > url.href
    => should be "git:/.//b", instead of "git://b"
   
    > const url = new URL("../b", "git:/.//a");
    > url.href
    => should be "git:/b", instead of "git:/./b"
   
    This CL also fixes the following case which replaces pathname with an
    empty path.
   
    > const url = new URL("git:/");
    > url.pathname = "";
    > url.href
    => should be "git:/", instead of "git:".
    > url.pathname
    => should be "/", instead of "".
   
    The URL Standard (https://url.spec.whatwg.org/#dom-url-pathname) defines
    the behavior of this case. Unfortunately, it is not easy to understand
    why url.pathname ends up as "/" in this case since the URL Standard
    appear cryptic. Please read the URL Standard carefully to understand
    that. I omit my understanding here.
   
    Tests:
   
    While the WPT URL tests provide good coverage for these scenarios, I
    also added some minimal unit tests for additional assurance.
   
    Bug: 40063064
    Change-Id: I036236218a0bfbe7be5ab147088dd6fd52a8212a
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5330846
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1273979}

M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element-xhtml_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/a-element_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-constructor.any_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters-a-area.window_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any.worker_exclude=(file_javascript_mailto)-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/url-setters.any_exclude=(file_javascript_mailto)-expected.txt
M       url/gurl_unittest.cc
M       url/url_canon.h
M       url/url_canon_non_special_url.cc
M       url/url_canon_relative.cc
M       url/url_canon_unittest.cc

https://chromium-review.googlesource.com/5330846

Project: chromium/src
Branch: main

commit 697a95529d7c01d00455e632af0deb5b7357b3be
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu Mar 21 05:53:25 2024

    Revert the change to use MakeRefCounted
   
    This is a follow-up of the prior review comment,
    https://crrev.com/c/5309015/comment/f1d24443_39524cab/.
   
    Bug: 40063064
    Change-Id: I21530feda6272ef550238f47428f25625565ef39
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5379437
    Auto-Submit: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1276062}

M       third_party/blink/renderer/platform/weborigin/security_origin.cc
M       third_party/blink/renderer/platform/weborigin/security_origin.h

https://chromium-review.googlesource.com/5379437

Project: chromium/src
Branch: main

commit 8ca53a2ad11fb434d772c08b711fa305e8b042a8
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Apr 01 03:40:26 2024

    Track the failing WPT tests in url-non-special virtual tests
   
    The following wpt tests will fail if we enable the
    kStandardCompliantNonSpecialSchemeURLParsing flag [^1].
   
    To address this, track these affected tests in url-non-special virtual
    tests, in addition to wpt/url tests, which are already being tracked.
   
    - external/wpt/fetch/data-urls/processing.any.html
    - external/wpt/fetch/data-urls/processing.any.serviceworker.html
    - external/wpt/fetch/data-urls/processing.any.sharedworker.html
    - external/wpt/fetch/data-urls/processing.any.worker.html
    - external/wpt/html/infrastructure/urls/resolving-urls/query-encoding/windows-1251.html
    - external/wpt/html/infrastructure/urls/resolving-urls/query-encoding/windows-1252.html
    - external/wpt/html/semantics/links/links-created-by-a-and-area-elements/non-special-url-getter-setter.window.html
    - virtual/keepalive-in-browser-migration/external/wpt/fetch/data-urls/processing.any.html
    - virtual/keepalive-in-browser-migration/external/wpt/fetch/data-urls/processing.any.serviceworker.html
    - virtual/keepalive-in-browser-migration/external/wpt/fetch/data-urls/processing.any.sharedworker.html
    - virtual/keepalive-in-browser-migration/external/wpt/fetch/data-urls/processing.any.worker.html
    - virtual/plz-dedicated-worker-disabled/external/wpt/fetch/data-urls/processing.any.html
    - virtual/plz-dedicated-worker-disabled/external/wpt/fetch/data-urls/processing.any.serviceworker.html
    - virtual/plz-dedicated-worker-disabled/external/wpt/fetch/data-urls/processing.any.sharedworker.html
    - virtual/plz-dedicated-worker-disabled/external/wpt/fetch/data-urls/processing.any.worker.html
    - virtual/produce-compile-hints/external/wpt/fetch/data-urls/processing.any.html
    - virtual/produce-compile-hints/external/wpt/fetch/data-urls/processing.any.serviceworker.html
    - virtual/produce-compile-hints/external/wpt/fetch/data-urls/processing.any.sharedworker.html
    - virtual/produce-compile-hints/external/wpt/fetch/data-urls/processing.any.worker.html
   
    [^1]: https://chromium-layout-test-archives.storage.googleapis.com/results.html?json=chromium/try/linux-rel/1752691/blink_wpt_tests%20%28with%20patch%29/full_results_jsonp.js
   
    Bug: 40063064
    Change-Id: I1c8ed4f670b549930f19750e303188d5543c9b63
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5399582
    Reviewed-by: Koji Ishii <kojii@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1280650}

M       third_party/blink/web_tests/VirtualTestSuites
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/fetch/data-urls/processing.any-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/fetch/data-urls/processing.any.serviceworker-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/fetch/data-urls/processing.any.sharedworker-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/fetch/data-urls/processing.any.worker-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/html/infrastructure/urls/resolving-urls/query-encoding/windows-1251_include=scheme-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/html/infrastructure/urls/resolving-urls/query-encoding/windows-1252_include=scheme-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/html/semantics/links/links-created-by-a-and-area-elements/non-special-url-getter-setter.window-expected.txt

https://chromium-review.googlesource.com/5399582

Project: chromium/src
Branch: main

commit 4c7f2eadc7b1f3cd652211eeeb46364b2ae998fa
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Apr 02 09:12:19 2024

    [URL] Track the failing web tests in url-non-special virtual tests
   
    The following web tests will fail if we enable the
    kStandardCompliantNonSpecialSchemeURLParsing flag [^1].
   
    To address this, track these affected tests in url-non-special virtual
    tests, in addition to wpt tests, which are already being tracked.
   
    - fast/dom/HTMLAnchorElement/set-href-attribute-hash.html
    - fast/dom/HTMLAnchorElement/set-href-attribute-hostname.html
    - fast/dom/HTMLAnchorElement/set-href-attribute-port.html
    - fast/dom/HTMLAnchorElement/set-href-attribute-search.html
    - fast/domurl/url-hash.html
    - fast/loader/javascript-url-hierarchical-execution.html
    - fast/loader/url-parse-1.html
    - fast/url/invalid-urls-utf8.html
    - fast/url/mailto.html
    - fast/url/segments.html
    - http/tests/security/base-url-data.html
    - http/tests/security/base-url-javascript.html
   
    [^1]: https://chromium-layout-test-archives.storage.googleapis.com/results.html?json=chromium/try/linux-rel/1752691/blink_web_tests%20%28with%20patch%29/full_results_jsonp.js
   
    Bug: 40063064
    Change-Id: I97a627da21436c3c57f44d51497fc9b7bad96e5c
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5401318
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Koji Ishii <kojii@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1281069}

M       third_party/blink/web_tests/VirtualTestSuites
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/dom/HTMLAnchorElement/set-href-attribute-hash-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/dom/HTMLAnchorElement/set-href-attribute-hostname-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/dom/HTMLAnchorElement/set-href-attribute-port-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/dom/HTMLAnchorElement/set-href-attribute-search-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/domurl/url-hash-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/loader/javascript-url-hierarchical-execution-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/loader/url-parse-1-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/invalid-urls-utf8-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/mailto-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/segments-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/http/tests/security/base-url-data-expected.txt
A       third_party/blink/web_tests/platform/linux/virtual/url-non-special/http/tests/security/base-url-javascript-expected.txt

https://chromium-review.googlesource.com/5401318

Project: chromium/src
Branch: main

commit 5d2acf1a08e3d68721f751bb2ee62ef186b656cf
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Apr 03 05:09:56 2024

    Remove wpt/svg/linking/reftests/url-processing-invalid-base.svg
   
    This test appears to be based on non-standard URL behavior. It assumes
    that a relative URL fragment like "#fragment" combined with the base URL
    "invalid:" results in an invalid URL. However, this is not standard
    compliant behavior. The combination of a relative fragment and the valid
    base URL should always result in a valid URL.
   
    Note: 'valid' (or 'invalid' ) in the context of this CL's description
    means what the URL Standard defines here:
   
    https://url.spec.whatwg.org/#valid-url-string
   
    In this sense, "invalid:" is a valid base URL.
   
    This test is currently passing on Chromium, but failing on Safari and
    Firefox.
   
    https://wpt.fyi/results/svg/linking/reftests/url-processing-invalid-base.svg?label=experimental&label=master&aligned
   
    If Chromium supports non-special URLs correctly
    (https://crbug.com/40063064), this test will fail on Chromium too.
   
    Since any valid base URL cannot be made invalid by any relative
    fragment, removing this test seems reasonable.
   
    This test was added in https://crrev.com/c/2115650.
   
    Bug: 40621475,40063064
    Change-Id: I46dee043ab02e7899a512e5a2cee3fc773848e5c
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5399999
    Reviewed-by: Fredrik Söderquist <fs@opera.com>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1281665}

D       third_party/blink/web_tests/external/wpt/svg/linking/reftests/url-processing-invalid-base.svg

https://chromium-review.googlesource.com/5399999

Project: chromium/src
Branch: main

commit 7d2d2baed2ad75a9555feddf4b752723935528b4
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Apr 09 07:28:25 2024

    [URL] Update url-hash.html test to handle non-special URLs correctly
   
    fast/domurl/url-hash.html is one of the tests which would fail if
    kStandardCompliandNonSpecialSchemeURLParser flag is enabled [^1].
   
    This test seems to test the current non-standard URL behavior in
    Chromium [^2].
   
    In preparation for shipping the feature, this CL does:
   
    - Update the test itself not to test the non-standard URL behavior.
    - Rebase the expectations of this test as failing and its
      corresponding virtual test in virtual/url-non-special as passing.
   
    When we ship the feature, the expectation of this test will be rebased
    as PASS.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
   
   
    Bug: 40063064
    Change-Id: I6930e3e340d276089c6de533c457c7ae6c47a5f0
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5433577
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1284320}

A       third_party/blink/web_tests/fast/domurl/url-hash-expected.txt
M       third_party/blink/web_tests/fast/domurl/url-hash.html
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/domurl/url-hash-expected.txt

https://chromium-review.googlesource.com/5433577

Project: chromium/src
Branch: main

commit 020b0367e57f90eadcaf73d756d7fde53ecebd03
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu Apr 11 08:36:10 2024

    [URL] Make invalid-urls-utf8.html test be standard compliant
   
    fast/url/invalid-urls-utf8.html is one of the tests which would fail
    if kStandardCompliandNonSpecialSchemeURLParser flag is enabled [^1].
   
    This test seems to test the current non-standard URL behavior in
    Chromium [^2].
   
    The test is using valid non-special URLs as invalid URLs, and made
    incorrect assmpptions about trimming behavior of `href` attribute
    valuie.
   
    In preparation for shipping the feature, this CL does:
   
    - Update the test to use only invalid URLs.
    - Remove expected.trim().
    - Add a test case for a special URL to ensure the value is not trimmed.
    - Rebase the expectations of this test as failing and its
      corresponding virtual test in virtual/url-non-special as passing.
   
    When we ship the feature, the expectation of this test will be rebased
    as PASS.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
   
   
    Bug: 40063064
    Change-Id: Ic61b6a3265faa71b028f21bead4e13ca3e22cc21
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5442113
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1285674}

M       third_party/blink/web_tests/fast/url/invalid-urls-utf8-expected.txt
M       third_party/blink/web_tests/fast/url/invalid-urls-utf8.html
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/invalid-urls-utf8-expected.txt

https://chromium-review.googlesource.com/5442113

Project: chromium/src
Branch: main

commit 77b1a5c737c23d2f6d881fbf14395c305f756613
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon Apr 15 02:36:39 2024

    [URL] Make mailto.html test be standard compliant
   
    fast/url/mailto.html is one of the tests which would fail
    if kStandardCompliandNonSpecialSchemeURLParser flag is enabled [^1].
   
    This test seems to test the current non-compliant URL behavior in
    Chromium [^2]. The following URL parser tester page might be helpful
    to understand the expected behavior:
   
    https://timothygu.me/urltester/#input=mailto%3Aa%2C+b
   
    In preparation for shipping the feature, this CL does:
   
    - Update the test to only test behavior that complies with the URL
      Standard.
    - Rebase the expectations of this test as failing and its
      corresponding virtual test in virtual/url-non-special as passing.
   
    When we ship the feature, the expectation of this test will be rebased
    as PASS.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
   
    Bug: 40063064
    Change-Id: I3e7bdcf2ffbbffd41fc521dc645ae41ae69a352d
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5444920
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1287129}

M       third_party/blink/web_tests/fast/url/mailto-expected.txt
M       third_party/blink/web_tests/fast/url/script-tests/mailto.js
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/mailto-expected.txt

https://chromium-review.googlesource.com/5444920

Project: chromium/src
Branch: main

commit ac4a61c77800d889c41012fab7bd8a0f4ca1ea5f
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Apr 16 08:51:49 2024

    [URL] Make base-url-data.html test be standard compliant
   
    http/tests/security/base-url-data.html is one of the tests which would
    fail if kStandardCompliandNonSpecialSchemeURLParser flag is enabled
    [^1] [^2].
   
    This test seems to test the current non-standard URL behavior in
    Chromium, which parses "data:/" URL as an opaque path URL, but that
    should be parsed as an path only URL. The following URL parser tester
    page might be helpful to understand the expected behavior:
   
    https://timothygu.me/urltester/#input=data%3A%2Fa+b
   
    In preparation for shipping the feature, this CL does:
   
    - Update the expected value of the test, which shouldn't affect the
      main purpose of the test.
    - Rebase the expectations of this test as failing and its
      corresponding virtual test in virtual/url-non-special as passing.
   
    When we ship the feature, the expectation of this test will be rebased
    as PASS.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
   
    Bug: 40063064
    Change-Id: Ic840bdee14918ee71ae41fac0d68762edd8a10ad
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5456558
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1287881}

M       third_party/blink/web_tests/http/tests/security/base-url-data-expected.txt
M       third_party/blink/web_tests/http/tests/security/base-url-data.html
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/http/tests/security/base-url-data-expected.txt

https://chromium-review.googlesource.com/5456558

Project: chromium/src
Branch: main

commit 5928ede6ec625d183ac8d38aa08de26bb1135787
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Apr 17 01:40:50 2024

    [URL] Don't navigate to an invalid "javascript:" scheme URL
   
    fast/loader/javascript-url-hierarchical-execution.html is one of the
    tests which would fail if kStandardCompliandNonSpecialSchemeURLParser
    flag is enabled [^1] [^2].
   
    It seems a browser navigates to "javascript://Spaceballs:" URL, and
    the navigation hits
    `DCHECK(!info.url_request.Url().ProtocolIs("javascript"))` [^3]:
   
    ```
    bool WebLocalFrameImpl::WillStartNavigation(const WebNavigationInfo& info) {
      DCHECK(!info.url_request.IsNull());
      DCHECK(!info.url_request.Url().ProtocolIs("javascript"));
      return GetFrame()->Loader().WillStartNavigation(info);
    }
    ```
   
    The test itself also fails because the custom policy delegate
    indicates that a load was attempted.
   
    Background / Investigation:
   
    1. When the feature is enabled, "javascript:" scheme URLs can be an
       invalid URL.
   
       e.g. "javascript://a b", which is a hierarchical non-special URL,
       but its host part contains a space, which is not allowed.
   
       Currently, without the flag, any "javascript:" scheme URL is
       considered as a valid URL because it is parsed as an *opaque path*
       URL.
   
    2. The current JavaScript execution code uses
       `KURL::ProtocolIsJavaScript()` to process a URL as "javascript:"
       URL, but it returns false for an invalid URL, even if the URL starts
       with "javascript:".
   
       So invalid "javascript:" URLs is not executed in a renderer and
       the URL is passed to navigation code. That upsets browser
       navigation code.
   
    Fix:
   
    I investigated the standard compliant behavior, and found
    wpt/url/javascript-urls.window.js. It seems the HTML and URL Standard
    requires that invalid "javascript:" URLs should NOT be executed.
   
    So the fix is to prevent a navigation for an invalid "javascript:"
    URL. See `FrameLoader::StartNavigation`.
   
    The CL also updates the description of
    `fast/loader/javascript-url-hierarchical-execution.html` to explain the intention of the test well. The previous description seems a bit outdated.
   
    Now both `fast/loader/javascript-url-hierarchical-execution.html` and
    `wpt/url/javascript-urls.window.js' tests in virtual/url-non-special
    pass. Neither JavaScript execution nor navigation happens for invalid
    "javascript:" URLs.
   
    Impacts:
   
    The following explanation is not directly related to this CL, but for
    the record, let me note the impact:
   
    The StandardCompliandNonSpecialSchemeURLParser feature affects URLs
    which start with "javascript://" or "javascript:/". Usually, such a
    "javascript:" URL contains only a comment. So there is no practical
    changes in most cases, unless a URL contains a new line ('%0a'),
    such as, "javascript://comment hello/%0aglobal_variable_a=1".
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
    [^3]: Why a browser receives "javascript://Spaceballs:", instead of
        "javascript://Spaceballs: The Comment!", is another interesting
        topic, but I omit the details in this CL. In short, invalid URL's
        string representation is not guaranteed.
   
    Bug: 40063064
    Change-Id: I914be8060166084815ba5186a106ecff00c179e9
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5454874
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1288459}

M       third_party/blink/renderer/core/loader/frame_loader.cc
M       third_party/blink/web_tests/fast/loader/javascript-url-hierarchical-execution-expected.txt
M       third_party/blink/web_tests/fast/loader/javascript-url-hierarchical-execution.html
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/external/wpt/url/javascript-urls.window-expected.txt
D       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/loader/javascript-url-hierarchical-execution-expected.txt

https://chromium-review.googlesource.com/5454874

Project: chromium/src
Branch: main

commit ab47a51f54e228d52c0e2961b71ca28d98d94bdd
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Apr 17 07:38:50 2024

    [URL] Make a failing Affiliation test pass with non-special URLs feature
   
    All/AffiliationUtilsMainDomainTest.ParamTest/14 is one of the tests
    which would fail when we enable the
    StandardCompliandNonSpecialSchemeURLParser feature [^1][^2]:
   
    The test is using a non-special URL,
    "android://blabla@com.twitter.android". When the feature is enabled, a
    non-special URL can be parsed correctly, which would affect the result
    of the test.
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by moving the test case into a parameterized test, in
    preparation for enabling the feature.
   
    I'd like to notify a code owner that we are changing the behavior of
    non-special URLs. Since I'm not familiar how Affiliation should work
    for non-special URLs, I'm leaving it to the original test author to
    verify or update the expected behavior, if necessary, possibly in
    another CL.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I086b02e7af4ea94a0c33961d5abef4c941a4c49d
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5459417
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Viktor Semeniuk <vsemeniuk@google.com>
    Cr-Commit-Position: refs/heads/main@{#1288553}

M       components/affiliations/core/browser/affiliation_utils_unittest.cc

https://chromium-review.googlesource.com/5459417

Project: chromium/src
Branch: main

commit 1d39f0cc405b24dd7a1deaf13d5228ab79376c93
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Apr 17 08:54:34 2024

    [URL] Update the expectation of fast/url/segments.html test
   
    fast/url/segments.html is one of the tests which would fail if
    kStandardCompliandNonSpecialSchemeURLParser flag is enabled [^1] [^2].
   
    This CL updates the expected values of the test, which are currently
    based on non-compliant behaviors of non-special URLs.
   
    I used `blink_tool.py rebaseline-cl' to rebase expectations across
    platforms.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I2dd4cc01d04e7e6eaff417a194a74c2016d68d8b
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5456994
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1288584}

M       third_party/blink/web_tests/fast/url/script-tests/segments.js
M       third_party/blink/web_tests/platform/linux/fast/url/segments-expected.txt
M       third_party/blink/web_tests/platform/linux/virtual/url-non-special/fast/url/segments-expected.txt
M       third_party/blink/web_tests/platform/mac/fast/url/segments-expected.txt
M       third_party/blink/web_tests/platform/win/fast/url/segments-expected.txt

https://chromium-review.googlesource.com/5456994

Project: chromium/src
Branch: main

commit 0346227cc62011c79c3a70c5e35e0243fb6785e2
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed Apr 17 09:38:57 2024

    [URL] Rewrite base-url-javascript.html and add variants
   
    http/tests/security/base-url-javascript.html is one of the tests which
    would fail if kStandardCompliandNonSpecialSchemeURLParser flag is
    enabled [^1] [^2].
   
    This test is using javascript URL, "javascript:// This is JavaScript",
    as a base URL. Currently, this is a valid JavaScript URL, however,
    once the StandardCompliandNonSpecialSchemeURLParser feature is
    enabled, this URL becomes an invalid URL because its host part
    contains a space character, which is not allowed. That's the reason
    this test would fail if the flag is enabled.
   
    For the comparison purposes, this CL rewrote and separated this test
    into three tests, as follows:
   
    1. base-url-https-invalid.html: Using an invalid https URL,
       "https://[xxx]".
    2. base-url-javascript-invalid.html: Using an invalid javascript URL,
       "javascript:// (space) " (hierarchical non-special URL).
    3. base-url-javascript-valid.html: Using a valid javascript URL,
       "javascript:..." (opaque path URL).
   
    Given that the current behaviors of Chromium is:
   
    - 1. If a base URL is invalid (in the sense of the URL Standard),
      image loading will fail, regardless of whether its scheme is "https"
      or "javascript".
    - 2. Otherwise, image should load successfully.
   
    The test result should be:
   
    |        | Flag off | Flag on |
    |--------+----------+---------|
    | Test 1 | PASS     | PASS    |
    | Test 2 | FAIL     | PASS    |
    | Test 3 | PASS     | PASS    |
   
    In preparation for shipping the feature, we mark the test 2 as FAIL,
    though all of the corresponding virtual/url-non-special tests are
    marked as PASS.
   
    When we ship the feature, all of the expectation will be rebased as
    PASS.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results?sortby=&groupby=
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/edit?usp=sharing
   
    Bug: 40063064
    Change-Id: Id7119d50e907961752e6b27119410c22a59f119b
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5456854
    Reviewed-by: Koji Ishii <kojii@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1288603}

M       third_party/blink/web_tests/VirtualTestSuites
A       third_party/blink/web_tests/http/tests/security/base-url-https-invalid.html
D       third_party/blink/web_tests/http/tests/security/base-url-javascript-expected.txt
A       third_party/blink/web_tests/http/tests/security/base-url-javascript-invalid-expected.txt
A       third_party/blink/web_tests/http/tests/security/base-url-javascript-invalid.html
M       third_party/blink/web_tests/http/tests/security/base-url-javascript-valid.html
D       third_party/blink/web_tests/platform/linux/virtual/url-non-special/http/tests/security/base-url-javascript-expected.txt
A       third_party/blink/web_tests/virtual/url-non-special/http/tests/security/base-url-javascript-invalid-expected.txt

https://chromium-review.googlesource.com/5456854

Project: chromium/src
Branch: main

commit de30b303934f59fbe6f6f7ff503d2d35fa13ae63
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu Apr 18 03:51:11 2024

    [URL] Use a valid javascript URL in PDFExtensionJSTest.Navigator tests
   
    The following tests would fail if we enable
    StandardCompliandNonSpecialSchemeURLParser feature [^1][^2]:
   
    - All/PDFExtensionJSTest.Navigator/0
    - All/PDFExtensionJSTest.Navigator/1
   
    The test currently uses a javascript URL,
    "javascript:// this is not a document.pdf", which will become an invalid
    URL when StandardCompliandNonSpecialSchemeURLParser flag is enabled.
   
    See the following URL parser tester for details:
    https://timothygu.me/urltester/#input=javascript%3A%2F%2F+this-is-not-a-document.pdf
   
    This CL updated the URL to a valid javascript URL,
    "javascript://this-is-not-a-document.pdf".
   
    Note: This test is not the only test which currently uses an invalid
    javascript URL. Several other tests previously using invalid
    JavaScript URLs have already been fixed.
   
    I'm leaving it to the original test author to decide whether to add a
    separate test for invalid URLs if that is necessary.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I8a8b3228216c7667a63db7c34d9467e89fb17c79
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5458678
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Lei Zhang <thestig@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1289131}

M       chrome/test/data/pdf/navigator_test.ts

https://chromium-review.googlesource.com/5458678

Project: chromium/src
Branch: main

commit c0ccf3b2ab3bfb32bcb50d627915dc5d48ac122c
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Apr 19 01:06:56 2024

    [URL] Make HistoryURLProviderTest test pass with non-special URL feature
   
    HistoryURLProviderTest.SuggestExactInput is one of the tests which
    would fail when we enable the
    StandardCompliandNonSpecialSchemeURLParser feature [^1][^2]:
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by making the test a parameterized test and adding
    expected outputs when the feature is enabled, in preparation for
    enabling the feature.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how HistoryURLProviderTest should work for
    non-special URLs (e.g. "mailto:" URL), I'm leaving it to the original
    test author to verify or update the expected behavior. A code owner
    might want to do a follow-up in another CL.
   
    I added several test cases for comparison purposes. Given that special
    URLs ("http:") and non-special URLs ("mailto:") have the same
    behavior, the expected outputs seem reasonable to me.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I3ec500f9b267599a877cb0eeea48e96b21d77705
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5465142
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Gang Wu <gangwu@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1289704}

M       components/omnibox/browser/history_url_provider_unittest.cc

https://chromium-review.googlesource.com/5465142

Project: chromium/src
Branch: main

commit 14089804e5b2e5e3fc6eb222b225999c2255fecd
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Apr 23 02:22:01 2024

    [URL] Make ProtocolHandlerRegistryTest pass with non-special URL feature
   
    ProtocolHandlerRegistryTest is one of the tests which would fail when
    we enable the StandardCompliandNonSpecialSchemeURLParser feature
    [^1][^2]:
   
    The test currently uses non-special URLs like::
   
    - 1. "web+custom://custom handler"
    - 2. "web+custom://custom/<>"
    - 3. "web+bool://user:password@example/"
   
    When the feature is enabled,
   
    - 1 is no longer a valid URL.
    - 2 is now correctly percent encoded.
      - '<': %3C
      - '>': %3E
    - The credential part of 3 is correctly parsed.
   
    That's why the test would fail with the feature enabled.
   
    See the following URL parser tester for details:
    1. https://timothygu.me/urltester/#input=web%2Bcustom%3A%2F%2Fcustom+handler
    2. https://timothygu.me/urltester/#input=web%2Bcustom%3A%2F%2Fcustom%2F%3C%3E
    3. https://timothygu.me/urltester/#input=web%2Bbool%3A%2F%2Fuser%3Apassword%40example%2F
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by making the test a parameterized test and adding
    expected outputs when the feature is enabled, in preparation for
    enabling the feature.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how ProtocolHandlerRegistryTest should work for
    non-special URLs (e.g. "web+custom://" URL here), it's recommended for
    a code owner or a domain expert to review and potentially do an
    appropriate action in a follow-up CL, if necessary.
   
    Depending on the severity of a potential breakage, I'd appreciate if
    you could file a bug and marking it as blocking for bug 40063064 so we
    don't ship the feature until all critical issues are addressed.
   
    Bug: 40063064
    Change-Id: I4668fcf088f836026d30fd8cd38d21bd48898319
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5467911
    Reviewed-by: Dominick Ng <dominickn@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1291033}

M       components/custom_handlers/protocol_handler_registry_unittest.cc

https://chromium-review.googlesource.com/5467911

Project: chromium/src
Branch: main

commit 1f1d0569a7fe28a05cfe5206319b3daa3e0f9f07
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue Apr 23 02:42:19 2024

    [URL] Use a valid non-special URL in AppSourceUrlRecorderTest test
   
    AppSourceUrlRecorderTest.CheckCrostini is one of the test which would
    fail if we enable the StandardCompliandNonSpecialSchemeURLParser
    feature [^1][^2]:
   
    The test currently uses a non-special URL,
   
    "app://I-💖.unicode!Und der Eisbär?/Name",
   
    which will become an invalid URL with the new feature enabled because
    it's host part includes a space character.
   
    See the following URL parser tester for details:
   
    https://timothygu.me/urltester/#input=app%3A%2F%2FI-%F0%9F%92%96.unicode%21Und+der+Eisb%C3%A4r%3F%2FName
   
    This CL updated the URL to a valid non-special "app:" URL,
    "app://I-💖.unicode!Und_der_Eisbär?/Name".
   
    I'd also like to notify a code owner that we are changing the behavior
    of non-special URLs via this CL's review. Since I'm not familiar with
    "app:" url, I'm leaving it to a code owner to decide whether to add a
    separate test for invalid "app:" URLs if that is necessary.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I1ee9905196751f04cab1fb2e63adc69f9edc728d
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5459056
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Jesse Doherty <jwd@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1291037}

M       components/ukm/app_source_url_recorder_test.cc

https://chromium-review.googlesource.com/5459056

Project: chromium/src
Branch: main

commit 395f7fe2af40dc3cec44f39552a18a5ae82ddb26
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu Apr 25 01:59:40 2024

    [URL] Make URLBlocklistManagerTest pass with non-special URL feature
   
    URLBlocklistManagerTest is one of the tests which would fail when we
    enable the StandardCompliandNonSpecialSchemeURLParser feature
    [^1][^2]:
   
    The test currently uses a non-special URL, like "about://newtab".
   
    When the feature is enabled, the host part in non-special URLs can be
    correctly parsed. That's why the test would fail with the feature
    enabled.
   
    See the following URL parser tester for details:
   
    https://timothygu.me/urltester/#input=about%3A%2F%2Fnewtab
    hostname:
    > whatwg-url : "newtab"
    > Chrome: (empty string)
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by making the test a parameterized test and adding
    expected outputs when the feature is enabled, in preparation for
    enabling the feature.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how URLBlocklistManager should work for non-special
    URLs (e.g. "about://", "chrome://" URLs here), it's recommended for a
    code owner or a domain expert to review and potentially do an
    appropriate action in a follow-up CL, if necessary.
   
    Depending on the severity of a potential breakage, I'd appreciate if
    you could file a bug and marking it as blocking for bug 40063064 so we
    don't ship the feature until all critical issues are addressed.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I63bf7c91462d1cf866e74b463b7326cca63a4467
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5474919
    Reviewed-by: Igor Ruvinov <igorruvinov@chromium.org>
    Reviewed-by: Julian Pastarmov <pastarmovj@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1292244}

M       components/policy/core/browser/url_blocklist_manager_unittest.cc

https://chromium-review.googlesource.com/5474919

Project: chromium/src
Branch: main

commit 3ab9b16c7c7713b2d90702e9e9581f13d7f7f46c
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu Apr 25 05:23:46 2024

    [URL] Make WebBundleUtilsTest pass with non-special URL feature
   
    WebBundleUtilsTest is one of the tests which would fail when we enable
    the StandardCompliandNonSpecialSchemeURLParser feature [^1][^2]:
   
    The test currently uses a non-special URL, like a
   
    "urn://uuid:00000000-0000-0000-0000-000000000000",
   
    which becomes an invalid URL when the feature is enabled.
   
    See the following URL parser tester for details:
    https://timothygu.me/urltester/#input=urn%3A%2F%2Fuuid%3A0-0
   
    This CL removes the usage of invalid non-special URLs from the test.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I8c2f6a2b50b5d1582e1d1e2f2805f8f93523f216
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5485969
    Reviewed-by: Adam Rice <ricea@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1292277}

M       components/web_package/web_bundle_utils_unittest.cc

https://chromium-review.googlesource.com/5485969

Project: chromium/src
Branch: main

commit 23ccc84f71d403539887be3c95c005e52178c63a
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri Apr 26 03:55:23 2024

    [URL] Make URLUtilTest pass with the non-special URL feature
   
    URLUtilTest is one of the tests which would fail when we
    enable the StandardCompliandNonSpecialSchemeURLParser feature
    [^1][^2]:
   
    The test currently uses a non-special URL, like "bogus://google.com".
   
    When the feature is enabled, the host part in non-special URLs can be
    correctly parsed. That's why the test would fail with the feature
    enabled.
   
    See the following URL parser tester for details:
   
    https://timothygu.me/urltester/#input=bogus%3A%2F%2Fgoogle.com3A%2F%2Fnewtab
    hostname:
    > whatwg-url : "google.com"
    > Chrome: (empty string)
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by making the test a parameterized test and adding
    expected outputs when the feature is enabled, in preparation for
    enabling the feature.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how `MatchFilters` should work for non-special URLs,
    it's recommended for a code owner or a domain expert to review and
    potentially do an appropriate action in a follow-up CL, if necessary.
   
    Depending on the severity of a potential breakage, I'd appreciate if
    you could file a bug and marking it as blocking for bug 40063064 so we
    don't ship the feature until all critical issues are addressed.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: Ia0f2b228f1f34a09b27a0c116ed1724c532977ed
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5475006
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Dominic Battre <battre@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1292858}

M       components/url_matcher/BUILD.gn
M       components/url_matcher/url_util_unittest.cc

https://chromium-review.googlesource.com/5475006

Project: chromium/src
Branch: main

commit de3dd2ae3e33463545e876e284ec393993cc0afa
Author: Hayato Ito <hayato@chromium.org>
Date:   Tue May 07 05:48:54 2024

    [URL] Make OfflinePageTest test pass with non-special URL feature
   
    OfflinePageHeaderTest.ToString is one of the tests which would fail
    when we enable the StandardCompliandNonSpecialSchemeURLParser feature
    [^1][^2]:
   
    The test currently uses a non-special URL,
   
    "content://foo/Bar \"\'\\Test"
   
    When the feature is enabled, space and double-quote characters within
    the path part of a non-special URL will be percent-encoded correctly.
    That's why the test would fail with the feature enabled.
   
    See the following URL parser tester for details:
   
    https://timothygu.me/urltester/#input=content%3A%2F%2Ffoo%2FBar+%22%27%5CTest
    > whatwg-url: content://foo/Bar%20%22'\Test
   
    This CL makes the test pass, regardless of whether the feature is
    enabled or not, by making the test a parameterized test and adding
    expected outputs when the feature is enabled, in preparation for
    enabling the feature.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how OfflinePageHeaderTest should work for non-special
    URLs (e.g. "content:" URL here), it's recommended for a code owner or
    a domain expert to review and potentially do an appropriate action in
    a follow-up CL, if necessary.
   
    Depending on the severity of a potential breakage, I'd appreciate if
    you could file a bug and marking it as blocking for bug 40063064 so we
    don't ship the feature until all critical issues are addressed.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I1610aba0b7dfc83b9a8b41e9c504d3bc80a19272
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5465286
    Reviewed-by: Justin DeWitt <dewittj@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1297296}

M       components/offline_pages/core/request_header/BUILD.gn
M       components/offline_pages/core/request_header/offline_page_header_unittest.cc

https://chromium-review.googlesource.com/5465286

Project: chromium/src
Branch: main

commit 5ed87db03083869b6ac60878a2410e3d03eb8b47
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon May 13 07:54:44 2024

    [URL] Opt-out "android:" scheme from non-special URL support
   
    Regarding standard compliant non-special URLs support, it
    turned out that we couldn't change the behavior of "android:" scheme
    URLs. I feel like that we have to introduce an escape hatch.
   
    See the discussion at https://crrev.com/c/5465607 for details.
   
    This CL introduces an escape hatch mechanism in SchemeRegistry to
    preserve the existing behaviors for specific non-special schemes,
    with "android://" being the sole example for now.
   
    At this moment, SchemeRegistry doesn't provide any mechanism to update
    the scheme list in runtime.
   
    As part of this change, I'm also reverting a prior CL:
    - https://crrev.com/c/5459417
   
    Bug: 40063064
    Change-Id: If8b8e6e3a601eb19169bb539332882cef4ee86a6
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5515685
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Reviewed-by: Viktor Semeniuk <vsemeniuk@google.com>
    Cr-Commit-Position: refs/heads/main@{#1299924}

M       components/affiliations/core/browser/affiliation_utils_unittest.cc
M       url/url_constants.h
M       url/url_util.cc
M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5515685

Project: chromium/src
Branch: main

commit 33c9f7fcf0871ac4dbc365010e015e17fcdcd708
Author: Hayato Ito <hayato@chromium.org>
Date:   Wed May 15 07:50:09 2024

    [URL] Move the function definition for better diff
   
    It's a preparatory step to improve the readability of the diff in
    https://crrev.com/c/5541313.
   
    Bug: 40063064
    Change-Id: I9574eddf25fb2a134b003d595e03e850fec80365
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5541196
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Commit-Queue: Daniel Cheng <dcheng@chromium.org>
    Auto-Submit: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1301150}

M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5541196

Project: chromium/src
Branch: main

commit ffa3455bacf5f95af59402b68d2c277ec9e8bb46
Author: Hayato Ito <hayato@chromium.org>
Date:   Thu May 16 02:06:31 2024

    [URL] Fix URLUtilTest tests for non-special scheme support
   
    Bug: 40063064
    Change-Id: Ie2c5d62290add5d92fdab7177de032104da66061
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5541313
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Daniel Cheng <dcheng@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1301714}

M       url/url_util_unittest.cc

https://chromium-review.googlesource.com/5541313

Project: chromium/src
Branch: main

commit d1f07e00d72e8f317e1aa7a50d5bcf0527f160a2
Author: Hayato Ito <hayato@chromium.org>
Date:   Fri May 17 01:44:04 2024

    [URL] Fix IsXCallbackURL() to support non-special scheme URL feature
   
    XCallbackURLTest.IsXCallbackURLtest is one of the tests which would
    fail [^1][^2] when we enable the
    StandardCompliandNonSpecialSchemeURLParser feature [^3]:
   
    The test will fail because, with the feature enabled,
    url.path_pierce() correctly return the path part of a non-special URL.
    Previously, it returns a wrong part.
   
    You can see the difference between the standard behavior and
    Chromium's current behavior here:
   
    https://timothygu.me/urltester/#input=git%3A%2F%2Fhost%2Fpath
   
    `url.pathname` returns "//host/path", instead of "/path", for
    "git://host/path" in current Chromium.
   
    This CL updates `IsXCallbackURL` to handle both scenarios. When the
    feature is enabled, it leverages the correct parsing result.
    Otherwise, it falls back to the previous (manual) parsing method. This
    ensures the test passes regardless of the feature state.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/ios-simulator/1877523/overview
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
    [^3]: http://bit.ly/url-non-special
   
    Bug: 40063064
    Change-Id: I4883410b12a0615e58e41784646cbfaa8ee242d2
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5541003
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Reviewed-by: Olivier Robin <olivierrobin@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1302366}

M       ios/chrome/common/BUILD.gn
M       ios/chrome/common/x_callback_url.cc
M       ios/chrome/common/x_callback_url_unittest.cc

https://chromium-review.googlesource.com/5541003

Project: chromium/src
Branch: main

commit d22f756ab3ba1ba5466ced3f10ba8eb7677db49a
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon May 20 03:30:41 2024

    [URL] Preserve the behavior of DatabaseIdentifier for non-special URLs
   
    DatabaseIdentifierTest.ExtractOriginDataFromIdentifier is one of the
    tests which would fail when we enable the
    StandardCompliandNonSpecialSchemeURLParser feature [^1][^2]:
   
    The test will fail because, with the feature enabled, url.host() now
    correctly returns the host part for non-special URLs. Previously, it
    always an empty string for non-special URLs.
   
    See http://bit.ly/url-non-special for details.
   
    This CL updates `DatabaseIdentifier::Parse` to preserve the current
    behavior to make the test pass regardless whether the feature is
    enabled or not.
   
    Though the main purpose of this CL is to make the test pass when the
    feature is enabled, I'd like to notify a code owner that we are
    changing the behavior of non-special URLs via this CL's review. Since
    I'm not familiar how DatabaseIdentifier should work for non-special
    URLs, it's recommended for a code owner or a domain expert to review
    and potentially do an appropriate action in a follow-up CL, if
    necessary.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1752691/test-results
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
   
    Bug: 40063064
    Change-Id: I410fba708ae0146703d3bb9690d3a1306cf4bb3b
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5526228
    Reviewed-by: Ayu Ishii <ayui@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1303194}

M       storage/common/database/database_identifier.cc
M       storage/common/database/database_identifier_unittest.cc

https://chromium-review.googlesource.com/5526228

Project: chromium/src
Branch: main

commit ddd96d726d2ffe780bfc2d9eae425e961ad91182
Author: Hayato Ito <hayato@chromium.org>
Date:   Mon May 20 04:35:17 2024

    [URL] Opt-out chromium-x-callback:// scheme from non-special URL support
   
    The following tests are failing [^1][^2] in ios-simulator bot when we
    enable the StandardCompliandNonSpecialSchemeURLParser feature [^3]:
   
    - AppStartupParametersTest.ParseURLWithChromeURLInXCallbackURL
    - AppStartupParametersTest.ParseURLWithJavascriptURLInXCallbackURL
    - AppStartupParametersTest.ParseURLWithMalformedXCallbackURL
    - AppStartupParametersTest.ParseURLWithXCallbackURL
    - AppStartupParametersTest.ParseURLWithXCallbackURLAndExtraParams
   
    It appears that these tests use "chromium-x-callback://" scheme.
   
    Usually, I would try to fix the failures somehow, however, iOS
    codebase presents a challenge for me (e.g. *.mm files, which scare
    me).
   
    Therefore, as a temporary solution, let me tentatively opt-out
    "chromium-x-callback://" from non-special URL support, as I did it for
    "android://" scheme [4]. That makes these failing tests pass
    regardless of the feature state.
   
    For a more permanent solution, code owners or domain experts familiar
    with iOS development might want to work on a real fix, if necessary.
    If a proper fix is critical before we ship the feature, I'd appreciate
    if you could file a bug and marking it blocking for bug 40063064 so
    we don't ship the feature until all critical issues are addressed.
   
    [^1]: https://ci.chromium.org/ui/p/chromium/builders/try/ios-simulator/1877523/overview
    [^2]: https://docs.google.com/spreadsheets/d/1Pqw1iKXK_lxHj-kLIAeRFs-khFz-BPZDio1W7SgEVE4/
    [^3]: http://bit.ly/url-non-special
    [^4]: https://crrev.com/c/5515685
   
    Bug: 40063064
    Change-Id: I17a0cd310059f8831d55304b6604f98133a0bdb1
    Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5539140
    Reviewed-by: Mark Cogan <marq@chromium.org>
    Commit-Queue: Hayato Ito <hayato@chromium.org>
    Cr-Commit-Position: refs/heads/main@{#1303205}

M       url/url_constants.h
M       url/url_util.cc

https://chromium-review.googlesource.com/5539140