AnnotationExtractor: fix tests on windows

[staabm]

fixes failing windows test

1) Rector\Tests\ChangesReporting\Annotation\AnnotationExtractorTest::testExtractAnnotationFromClass with data set "Class with changelog annotation" ('Rector\Tests\ChangesReporting...ngelog', '@changelog', 'rectorphp/...iew.md')
Failed asserting that two strings are identical.
--- Expected
+++ Actual
@@ @@
-'rectorphp/rector@master/docs/rector_rules_overview.md'
+'rectorphp/rector@master/docs/rector_rules_overview.md
'

2) Rector\Tests\ChangesReporting\Annotation\AppliedRectorsChangelogResolver\RectorsChangelogResolverTest::test
Failed asserting that two arrays are identical.
--- Expected
+++ Actual
@@ @@
 Array &0 (
-    'Rector\Tests\ChangesReporting\Annotation\AppliedRectorsChangelogResolver\Source\RectorWithChangelog' => 'rectorphp/rector@master/docs/rector_rules_overview.md'
+    'Rector\Tests\ChangesReporting\Annotation\AppliedRectorsChangelogResolver\Source\RectorWithChangelog' => 'https://github.com/rectorphp/rector/blob/master/docs/rector_rules_overview.md\r
+'
 )

[staabm]

ready to land

[TomasVotruba]

What exactly is going on here?

This looks like a workaround for a bug in regex. We should fix the bug in regex instead. Ideally with link to verification on https://regex101.com/

[staabm]

It seems the regex is matching \r\n and therefore the \r\n is part of the matched url at the very end

Actually it could be reduced to a rtrim. I see no easy way to support a multi line regex fixing the issue at hand.

[TomasVotruba]

If rtrim() does the same work, let's go for it 👍

[staabm]

Fix verified. Ready to land

[staabm]

regarding the regex.

question is, why does https://3v4l.org/nU270 match the string without a ending newline, but https://3v4l.org/G9NNG contains the ending newline in the matched string

[staabm]

fixed the regex and reverted the rtrim workaround. ready to land.

[staabm]

@TomasVotruba good to go

[TomasVotruba]

Thank you 👍 Elegant fix