487192 - Move both copies (one is char, the other is PRUnichar) of nsWildCard into xpcom

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[jag (Peter Annema)]

We've got two copies of nsWildCard.{h,cpp}, one for |char*| and one for |PRUnichar*|. Some fixes and license changes were only applied to the |char*| version, and I'd like to address that by using template functions for the implementation code (wrapped inside / hidden by the publicly visible functions) so any future fixes only need to be made once, and I'd like to move the result into xpcom/io/ or some other suitable place.

Comment 1

[(not currently active) Ted Mielczarek]

See also bug 484353.

Comment 2

[jag (Peter Annema)]

Attachment: Step 1: Sync up filepicker's nsWildCard to fixes made to libjar's version.

Comment 3

[jag (Peter Annema)]

Attachment: Step 2: Clean up memory usage. Use templates for helpers. Make both copies equal.

Comment 4

[jag (Peter Annema)]

Attachment: Step 3: Move nsWildCard.{h,cpp} into xpcom/io and start using it.

Comment 5

[jag (Peter Annema)]

Attachment: Step 4: Whitespace clean-up

Comment 6

[(not currently active) Ted Mielczarek]

You know I'm not actually an XPCOM peer, right?

Comment 7

[jag (Peter Annema)]

Hrm.

Comment 8

[jag (Peter Annema)]

http://hg.mozilla.org/mozilla-central/rev/3251f66498f9
http://hg.mozilla.org/mozilla-central/rev/4f0c78fb34fc
http://hg.mozilla.org/mozilla-central/rev/d1f45c237fc4
http://hg.mozilla.org/mozilla-central/rev/5eeacfad0e2b

Comment 9

[jag (Peter Annema)]

Bustage fix:

http://hg.mozilla.org/mozilla-central/rev/4e91deae75dc

Comment 10

[Mike Beltzner [:beltzner, not reading bugmail]]

Regression: DHTML increase 1.43% on XP Firefox
http://graphs-new.mozilla.org/graph.html#tests=[{%22machine%22:55,%22test%22:18,%22branch%22:1},{%22machine%22:56,%22test%22:18,%22branch%22:1},{%22machine%22:57,%22test%22:18,%22branch%22:1},{%22machine%22:71,%22test%22:18,%22branch%22:1}]&sel=1242197940,1242370740

Previous results: 1073.91 from build 20090513234400 of revision 5eeacfad0e2b at 2009-05-13 23:44:00 on qm-pxp-fast03 run # 0

New results: 1089.29 from build 20090513235938 of revision 4e91deae75dc at 2009-05-13 23:59:00 on qm-pxp-trunk03 run # 0

Suspected checkin range: from revision 5eeacfad0e2b to revision 4e91deae75dc
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=5eeacfad0e2b&tochange=4e91deae75dc

... implicates this patch

Comment 11

[jag (Peter Annema)]

Highly unlikely that these changes affected DHTML.

Also, have you seen the graphs you linked to? 1) They've been steadily increasing over the past day. 2) There's a big dip around the time you quote. I'm not sure what caused the dip, but if you ignore that, the difference is negligible and in line with the trend.

Now, I'm not saying we shouldn't do anything here, that increasing trend should be explained, and so should the dip, but I'm reluctant to back this out unless you can give me something a bit more solid.