Improve Character emulation #10113
Improve Character emulation #10113
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
zbynek
force-pushed
the
character-emul
branch
6 times, most recently
from
051a0d0 to
1ab960c
Compare
zbynek
force-pushed
the
character-emul
branch
2 times, most recently
from
d3a8f64 to
1ad6bc2
Compare
| // Known differences between Java 17 and Chrome 135 | ||
| // 11f50 .. 11f59, 16ac0 .. 16ac9, 1e4f0 .. 1e4f9, 1fbf0 .. 1fbf9 |
There was a problem hiding this comment.
Should we leave this method out if it is still wrong? Which side do we consider to be "wrong" here, Java or Chrome? Would it make sense to have a test (possibly ignored) that shows this, so it is easier to reevaluate later?
Also, consider using hex in digit() above, so that it is the same convention as here, or change this to be decimal?
There was a problem hiding this comment.
Maybe instead of this comment I should just link to the compatibility table https://docs.oracle.com/en/java/javase/24/docs/api/java.base/java/lang/Character.html#conformance
and mention that JRE behaviour on any Java < 24 won't match recent browser releases.
|
|
||
| /** | ||
| * Tests for java.lang.Character. | ||
| */ | ||
| @DoNotRunWith(Platform.HtmlUnitBug) |
There was a problem hiding this comment.
Can the specific failing tests be annotated with this, each with a comment about why it fails, rather than skip existing tests that pass?
There was a problem hiding this comment.
I's pretty much all the tests that call any Character.is* method. They fail because the regexp doesn't even parse in HtmlUnit 2-4. If we want any meaningful test coverage we need #10115 + a custom build of HtmlUnit that includes mozilla/rhino#1848
| } | ||
|
|
||
| // If String.toUpperCase produces more than 1 codepoint, Character.toUpperCase should | ||
| // act either as identity or title-case conversion (not supported in GWT). |
There was a problem hiding this comment.
Maybe add a failing test and mark it ignored so we can see about restoring it when it works?
Co-authored-by: Colin Alworth <[email protected]>
Fixes #9705
Fixes #1989 (not 100% of methods are covered, but some are just impossible without bundling unicode database into compiled code)
Implements several APIs using https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Regular_expressions/Unicode_character_class_escape that should be well supported in all browsers.
The implementation is mostly similar to https://groups.google.com/g/google-web-toolkit-contributors/c/73-aScAShs4/m/gZAhlUXiBAAJ , but the fallbacks for old Edge versions are not included.