Closed Bug 42280 Opened 26 years ago Closed 25 years ago

post-RTM, should replace all instances of http://www.w3.org/TR/REC-html40 URI in our XUL with correct XHTML URI

Categories

(Core :: XUL, defect, P3)

All
Windows NT
defect

Tracking

()

RESOLVED FIXED
Future

People

(Reporter: ekrock, Assigned: trudelle)

References

Details

(Whiteboard: Is there anything left to do for mozilla 1.0 ?)

[This is a post-RTM cleanup issue; opening this bug for tracking purposes.] Currently, the XUL code in our product in many places/files declares http://www.w3.org/TR/REC-html40 as the URI for the namespace enabling the use of HTML within XML. The URI should really be http://www.w3.org/1999/xhtml instead. For now, both work within our product (see bug 7834 for discussion), so we needn't change this for RTM (and indeed can't afford to, as we don't have the luxury of such cleanup at this point), but long-term, we don't want all our product XUL to be encouraging people to do the Wrong Thing, so we want to fix this post-RTM. Once *this* bug is fixed and our product is 100% free of REC-html40, we *may* choose to officially drop REC-html40 support from the product in a future release. We'll be officially warning developers about this through a TechNote on mozilla.org that dbaron is writing.
Marking FUTURE, and noting this this bug blocks bug 7834, "REC-html40 should not be an HTML namespace".
Blocks: 7834
Target Milestone: --- → Future
Status: NEW → ASSIGNED
rbs whacked the tree to change all the xmlns: references in .xul files to use http://www.w3.org/1999/xhtml early on jun 15. Two usages have crept back in however: ./extensions/psm-glue/res/content/securityOverlay.xul ./xpfe/communicator/resources/content/dialogBindings.xml Perhaps someone should make an announcement that the use of 'http://www.w3.org/TR/REC-html40' is deprecated in new XUL files. (In addition, Eric|Vidur, you probably need to file a bug in bugscape to get any files in the commercial tree changed over to the new value)
[mid-air collision..., resubmitting] Clean-up is now complete (I changed all the occurrences in the code tree and the doc tree). If I had access to the commercial tree, I could clean there as well :-) Are the files similar? Is synch:ing between the two trees possible? The support for REC-html40 that was left as a precaution is still lurking in nsNameSpaceManager.cpp. The side-effect of this is that, due to the old habit, people can erroneously continue to use REC-html40 because they will test and see that every thing is alright. As jrgm pointed out, this is already happening. And if it goes on like this, we may be back to square one pretty soon. What about removing (commenting-out) this support now that no problems have been reported with the clean-up, or is it better to do it post-beta2? Are other problems anticipated?
If you send me the perl script, I might be able to do it on the commercial tree files (which are mostly *additional* files rather than replacements for mozilla tree ones, I think). The problems it would cause would be that chrome designed by others would suddenly break, which might not be the best thing to do right before beta (since it wouldn't give people much warning).
David, I am behind the firewall now, without access to files. But I will be sending you the perl script once I have access to it. It needs to be run thrice: 1: with just_show_me set (i.e., no replacements yet, verification) 2: with just_show_me unset (i.e., go ahead, do the replacements) 3: with checkin set, (i.e., go ahead, do the checkins of the changed files, while prompting to give the opportunity to skip/abort/continue -- For example, I used this to skip some MathML files because I had some ongoing changes that I didn't want to checkin right at the time of the replacements). If the files are local files, then the last run (for checkins) is not needed. And so, Eric/Vidur, the perl script might also help to clean-up your demos. It might also be a good idea to send the perl script to skin designers. Since usage is straightforward, ('replace.pl -i <old> <new> <dir>'), they would have a greater incentive to sync themselves with Mozilla. In a sense, this would amount to the work (at a much lower-scale) that would have been done to advise them not to use REC-html40. When will nsbeta2 go out? in a month? On the fast Mozilla scale, this is plenty of time for them to type 'replace.pl -i <old> <new> <dir>', don't you think so? Anyway, I do not have a strong preference either way because I don't know all the surrounding parameters. I was just concerned that the usage of the old REC-html40 was spreading too fast, and would spread even faster with the release of nsbeta2.
Just forwarded the perl script to dbaron.
The commercial tree is now free of REC-html40. (I actually used a 1-line sed script run by a 4-line shell script instead of messing with perl.)
David. Can we close this bug out now? Also, should you (or vidur or someone) post to .xpfe,.ui,.layout,[.seamonkey] to let everyone know that they should use the official namespace in all new XUL files.
David has also been careful to clean-up the other REC-html40 that crept back. It is relieving to know that all existing trees are now clean. This sounds like a marvellous opportunity to comment-out the support of REC-html40...
Nominating for mozilla 0.8 to trigger an evaluation of the status of this bug. If there are any issues left, they should be resolved for mozilla 1.0, I guess.
Keywords: mozilla0.8
Whiteboard: Is there anything left to do for mozilla 1.0 ?
Sounds like there is nothing left to do here. Does someone want to take it and resolve as fixed, or should I mark WFM?
I looked at lxr for mozilla and commercial trees, and there are no xul files that use the REC-html40 namespace URI. Any uses that do creep in in the future can be treated as separate bugs. Marking this bug as FIXED.
Status: ASSIGNED → RESOLVED
Closed: 25 years ago
Resolution: --- → FIXED
Two questions: 1. What about this one? /extensions/xmlterm/ui/content/xmltermOverlay.xul, line 8 -- xmlns:html="http://www.w3.org/TR/REC-html40" 2. What about rbs' suggestion about commenting out support for REC-html40 in nsNameSpaceManager.cpp to avoid regressions?
> 1. What about this one? /extensions/xmlterm/ui/content/xmltermOverlay.xul Good catch. (My bad. I don't know how I missed that ...) I filed bug 65662 for that instance. > 2. What about rbs' suggestion about commenting out support for REC-html40 > in nsNameSpaceManager.cpp to avoid regressions? I don't believe we can arbitarily drop support for this, since there are more .xul files than just those in the mozilla trees. But that would best be discussed in the newsgroups and a new bug filed for whatever action should be taken.
Keywords: mozilla0.8
Depends on: 155506
No longer depends on: 155506
Component: XP Toolkit/Widgets: XUL → XUL
QA Contact: jrgmorrison → xptoolkit.widgets
You need to log in before you can comment on or make changes to this bug.