1
0
Fork 0
mirror of https://we.phorge.it/source/phorge.git synced 2024-11-19 13:22:42 +01:00

Markup project hashtags which begin with (or contain only) digits

Summary:
Fixes T9832. We currently refuse to recognize project hashtags in remarkup if they begin with a digit. This is motivated by attempting to not recognize them if they contain only digits.

I don't think we really gain anything by this. Although most `#123` in text are probably not project references, the cost of doing a lookup for them is quite small, and //some// of them are.

In cases where users use `#123` to refer to tasks in an external system, they can use a rule for that with higher precedence than this one or not give their projects conflicting hashtags.

Test Plan:
  - This is well-covered by unit tests.
  - Referenced a `#3u1`, per T9832.

Reviewers: chad

Reviewed By: chad

Maniphest Tasks: T9832

Differential Revision: https://secure.phabricator.com/D14547
This commit is contained in:
epriestley 2015-11-23 05:51:14 -08:00
parent 2a063a93a9
commit 2f7010d18e
2 changed files with 24 additions and 17 deletions

View file

@ -20,17 +20,13 @@ final class ProjectRemarkupRule extends PhabricatorObjectRemarkupRule {
}
protected function getObjectIDPattern() {
// NOTE: This explicitly does not match strings which contain only
// digits, because digit strings like "#123" are used to reference tasks at
// Facebook and are somewhat conventional in general.
// The latter half of this rule matches monograms with internal periods,
// like `#domain.com`, but does not match monograms with terminal periods,
// because they're probably just puncutation.
// NOTE: The latter half of this rule matches monograms with internal
// periods, like `#domain.com`, but does not match monograms with terminal
// periods, because they're probably just puncutation.
// Broadly, this will not match every possible project monogram, and we
// accept some false negatives -- like `#1` or `#dot.` -- in order to avoid
// a bunch of false positives on general use of the `#` character.
// accept some false negatives -- like `#dot.` -- in order to avoid a bunch
// of false positives on general use of the `#` character.
// In other contexts, the PhabricatorProjectProjectPHIDType pattern is
// controlling and these names should parse correctly.
@ -38,17 +34,14 @@ final class ProjectRemarkupRule extends PhabricatorObjectRemarkupRule {
// These characters may never appear anywhere in a hashtag.
$never = '\s?!,:;{}#\\(\\)"\'';
// These characters may not appear at the beginning.
$never_first = '.\d';
// These characters may not appear at the end.
$never_last = '.';
// These characters may not appear at the edge of the string.
$never_edge = '.';
return
'[^'.$never_first.$never.']+'.
'[^'.$never_edge.$never.']+'.
'(?:'.
'[^'.$never.']*'.
'[^'.$never_last.$never.']+'.
'[^'.$never_edge.$never.']+'.
')*';
}

View file

@ -33,7 +33,21 @@ final class ProjectRemarkupRuleTestCase extends PhabricatorTestCase {
),
'#123' => array(
'embed' => array(),
'ref' => array(),
'ref' => array(
array(
'offset' => 1,
'id' => '123',
),
),
),
'#2x4' => array(
'embed' => array(),
'ref' => array(
array(
'offset' => 1,
'id' => '2x4',
),
),
),
'#security#123' => array(
'embed' => array(),