Closed
Bug 7834
Opened 26 years ago
Closed 23 years ago
REC-html40 should not be an HTML namespace
Categories
(Core :: XML, defect, P3)
Core
XML
Tracking
()
RESOLVED
FIXED
Future
People
(Reporter: dbaron, Assigned: nisheeth_mozilla)
References
()
Details
Attachments
(10 files)
Right now you only seem to support the URI http://www.w3.org/TR/REC-html40 for
the html namespace. (It is now used in the above document.) However, I think
you should support the ones in XHTML (and track it as it becomes a REC), since
those are the only formally defined namespaces. When the namespace URI was:
http://www.w3.org/TR/xhtml1
then the above document didn't work (which is also the link to the spec).
Reporter | ||
Updated•26 years ago
|
Assignee: trudelle → nisheeth
Component: XUL → XML
Reporter | ||
Comment 1•26 years ago
|
||
Whoops. Accidentally hit wrong module. One letter *does* make a difference.
Assignee | ||
Updated•26 years ago
|
Status: NEW → ASSIGNED
Target Milestone: M9
Assignee | ||
Comment 2•26 years ago
|
||
Moving this to M9 when we will do a lot of the HTML/XML/XUL content sink code
factoring. Support for XHTML might be an easy fallout of that effort.
Assignee | ||
Updated•26 years ago
|
Summary: XHTML namespaces should be supported → [Content sink] XHTML namespaces should be supported
Assignee | ||
Comment 3•26 years ago
|
||
CCing Vidur so that he is aware of this bug. He is going to do the content sink
factoring.
Assignee | ||
Comment 4•26 years ago
|
||
The content sink cleanup work is probably going to happen in M10. Moving this
bug accordingly.
Assignee | ||
Updated•26 years ago
|
Target Milestone: M9 → M10
Assignee | ||
Updated•25 years ago
|
Target Milestone: M10 → M12
Assignee | ||
Comment 5•25 years ago
|
||
Moving content sink code factoring related bugs to M12.
Assignee | ||
Updated•25 years ago
|
Target Milestone: M12 → M15
Assignee | ||
Comment 6•25 years ago
|
||
The content sink code factoring has been postponed to post beta. Moving related
bugs out to M15.
Updated•25 years ago
|
QA Contact: ckritzer → chrisd
Reporter | ||
Comment 7•25 years ago
|
||
What's the status of this bug? I think this should be relatively easy to do,
right? XHTML became a recommendation today, and the official namespace URI is:
http://www.w3.org/1999/xhtml
Frankly, if you're going to claim that Mozilla's XML support is standards-based,
you have to do this, since right now the only way to use the HTML vocabulary in
XML documents in to use the non-standard http://www.w3.org/TR/REC-html40 URI.
However, it does bring up a few tricky issues in the style system (supporting
multiple namespaces for @namespace) and other places **UNLESS** you completely
remove the support for the nonstandard http://www.w3.org/TR/REC-html40 (which I
believe is a pure Mozilla-ism, except that it was used in a few made up examples
in the xml-names REC which are probably non-normative). This would be an easy
change - a global search-replace for the HTML namespace URI in:
* the namespace code that recognizes it (html.css and some C++ code)
* all the XUL files that use it
* any sample code written using the nonstandard way. (?)
Assignee | ||
Comment 8•25 years ago
|
||
I'm going to bring in this bug to M14. It'll be nice to get this done by beta
and I'll try hard to make it happen...
Target Milestone: M15 → M14
Comment 9•25 years ago
|
||
XBL is another case of re-entrant parsing. This seems to work fine on the tip
before your changes. :)
What's the assertion?
Assignee | ||
Comment 10•25 years ago
|
||
These XML bugs need to move to M15, unfortunately. These aren't beta 1 stoppers
and I have to pull in other more critical bugs from Kipp's plate for beta 1.
Target Milestone: M14 → M15
Comment 11•25 years ago
|
||
I'd like to see this bug fixed soon, please. It affects mathml. In particular
all latex to mathml convertors give a xhtml namespace not this html4.0
namespace. Of course this can be fixed by hand... But this is inelegant, and
gives non-standard xml/mathml pages.
Comment 12•25 years ago
|
||
The minimal fix to support the XHTML namespace in the XML content sink is
included below. I'd like to see if it's possible to include this for beta1. It
would bolster our standards support (even thought it's not a complete
implementation) and is a very low risk fix for a potentially high gain. I
believe Eric Krock agrees.
Index: nsNameSpaceManager.cpp
===================================================================
RCS file: /cvsroot/mozilla/layout/base/src/nsNameSpaceManager.cpp,v
retrieving revision 3.9
diff -r3.9 nsNameSpaceManager.cpp
36a37
> static const char kXHTMLNameSpaceURI[] = "http://www.w3.org/1999/xhtml";
111a113
> nsString* xhtml = new nsString(kXHTMLNameSpaceURI);
114a117
> gURIArray->AppendElement(xhtml);
117a121
> NameSpaceURIKey xhtmlKey(xhtml);
120a125
> gURIToIDTable->Put(&xhtmlKey, (void*)kNameSpaceID_HTML);
Keywords: beta1
Comment 13•25 years ago
|
||
I agree that this is low risk and a valuable feature.
Comment 15•25 years ago
|
||
Comment 16•25 years ago
|
||
Oh, crap, I attached that testcase to the wrong bug! Sorry about the spam.
Comment 17•25 years ago
|
||
Comment 18•25 years ago
|
||
Reporter | ||
Comment 19•25 years ago
|
||
There's only one official HTML namespace -- the XHTML one. Using REC-html40 as
a namespace URI should go away.
Assignee | ||
Comment 20•25 years ago
|
||
Updating summary...
Summary: [Content sink] XHTML namespaces should be supported → REC-html40 should not be an HTML namespace
Assignee | ||
Comment 22•25 years ago
|
||
We are not going to try and remove support for REC-html40 before beta2.
Setting milestone to M19...
Target Milestone: M16 → M19
Assignee | ||
Comment 23•25 years ago
|
||
This bug has been marked "future" because the original netscape engineer working
on this is over-burdened. If you feel this is an
error, that you or another known resource will be working on this bug,or if it
blocks your work in some way -- please attach your
concern to the bug for reconsideration.
Target Milestone: M19 → Future
Comment 24•25 years ago
|
||
There is absolutely _no_way_ that we can let Netscape ship a browser which
recognises http://www.w3.org/TR/REC-html40 as a namespace. That would make it
completely impossible for people to author valid XHTML documents.
It is absolutely *imperative* that by FCS we use the correct XHTML namespace:
http://www.w3.org/1999/xhtml
...otherwise we will basically be killing the work of the XHTML WG in deciding
what the namespace is.
Removing Future milestone. This really should have been fixed yesterday (beta1)
so that content developers do not get annoyed when we finally do change.
Nominating for nsbeta2.
Comment 25•25 years ago
|
||
Mozilla has spent a lot of work adding XML support. XHTML is not only the
current HTML standard but is essential for the current modularization effor for
HTML. The VAST majority of XML applications will be mixing various XML
document types (SVG, MathML, XHTML modules, etc) within XML/XHTML documents.
It is arguable that without proper XHTML (and general) namespace support, any
support Mozilla does have for XML/XHTML is basically nullified. What good is
supporting XML/XHTML if you can't display any usefull/real documents? I don't
consider Mozilla as even supporting XML until this bug is fixed.
Reporter | ||
Comment 26•25 years ago
|
||
Just to clarify, right now I think mozilla supports both namespaces via some
strange mapping. However, I think most of the documents (e.g., XUL) use the
wrong one (which sets a bad example), so the first step is to change all that
stuff over to the right thing...
Comment 27•25 years ago
|
||
I should have tested again before I spoke, as I hadn't actually tested this bug
in a while. David is right, unlike my previous attempt, both namespaces are now
mapped to the same "value". The last time I tried it wouldn't even recognize
the proper xhtml namespace, only the improper HTML one. The behavior as it
stands is at least useable.
That brings us to the core issue with this bug. I do not have a problem with
recognizing (applying semantics to) the Rec-html40 namespace. I question the
current behavior of merging the "attributes" of both the HTML and XHTML
namespaces into one. I would think that the proper behavior, as the third
attachment above demonstrates, would be for the semantics of both namespaces to
be recognized, but "attributes" kept in the separate namespaces as they are
declared, not merged into one "pseudo" space. As in the attachment, I think
you should be able to style HTML:H1 and XHTML:H1 independently, being as they
are declared to be in *different* namespaces, albeit those namespaces have the
same semantics.
Comment 28•25 years ago
|
||
Putting on [NEED INFO] radar. nisheeth, please speak with rickg.
Whiteboard: [NEED INFO]
Assignee | ||
Comment 29•25 years ago
|
||
I think that supporting HTML semantics for elements declared in the REC-HTML40
namespace is only neeeded for supporting the internal XUL files that use
this functionality currently. I don't think the web community is going to use
that namespace, now that we have the XHTML namespace to do the same thing. So,
the intent behind "futuring" this bug was the desire to avoid spending time on
cleaning up the internal XUL files to use the new XHTML namespace.
But, I see your point, Chris, that people might want to style REC-HTML40
namespaced elements/attributes. At the same time, I differ from your opinion in
that I don't think that the REC-HTML40 namespace should have any HTML semantics
associated with it. The XHTML namespace is now the standards based way of
associating HTML semantics in an XML document. We should not confuse the web
community by supporting REC-HTML40 also.
In my opinion, the work to remove RECHTML40 support is not a beta 2 blocker. I
think the earliest we should try to do this is for beta 3. Ian, if you agree
with this stance, please remove the nsbeta2 keyword. I am adding the nsbeta3
keyword.
Keywords: nsbeta3
Comment 30•25 years ago
|
||
Actually, after considering it, I withdraw my statement about not having a
problem with applying semantics to REC-html4. I had being playing with
namespaces in Moz and IE5, and I guess what led me to stating I was not dead
set againsed it was that it is the _only_ "html semantic" namespace recognized
by IE5, and thus the only _current_ cross browser method of applying those
semantics to an XML document. But you are right, supporting it will just lead
to more confusion, and even worse (because of the cross browser thing), people
may actually use it. I don't so much mind stuff like backward compatibility
with the old HTML DOM, because the HTML document base is already a complete
mess, but I would like to see XML/XHTML documents stay nice (style free), shiny
(well-formed/valid), and very compliant with standards (no extraneous
namespaces :-)
Comment 31•25 years ago
|
||
Chris, if REC-html4 is the _only_ "html semantic" namespace recognized
by IE5, and thus the only _current_ cross browser method of applying those
semantics to an XML document, and we drop support for it, how will web content &
web app developers create cross-browser content? Please explain.
Here's my take:
1) We apparently already make it possible for developers to Do The Right Thing
(use the XHTML namespace) if they are so inclined, and it will work. That's the
critical thing for FCS. We can educate developers to use the Right Way.
2) Making it impossible to do the wrong thing, while in theory desirable, is not
necessarily a requirement for FCS. Caveat emptor.
3) If it's true that using the REC-html4 is the *only* way to use HTML elements
in XML on IE, disabling REC-html4 may well be unwise and make life harder for
content developers. As much as I would like to be able to dictate that all
content developers Do The Right Thing and that all browser vendors Support The
Right Thing, the fact is that I can't: IE5 has substantial market share and
limited standards support, and the unfortunate reality is that it's not going to
disappear when N6/Moz ships. Content developers will be trying to write for
both.
If we've given them the ability to do things Right when they want to and not
work on IE (and it appears that we have), then we shouldn't necessarily take
away the option of doing things Less Right and working on both IE5 and N6/Moz.
That's a choice for the content developers to make.
Chris, if you can explain a way that content developers can equally easily
create working cross-browser XML with HTML elements for IE5 and N6/Moz if we
disable REC-html40, then it makes sense to consider doing so, but if not, I
think we'd better leave the option in.
BTW, I'm the first to admit I don't fully understand all the issues here, so
folks please educate me (perhaps including simple sample markup examples!) if
you've got the time. Thanks!!!
Comment 32•25 years ago
|
||
As far as cross browser considerations, as a content developer, I am totally
frustrated for a million reasons already - I want to stick my head in the sand,
code to the specs, and let the browser manufacturers sort out the mess. You
could let users Do The Wrong Thing if they must, be assured that it is this
decision right here that very well could lead to a world full of REC-html4
namespaced documents, with nobody following the proper spec. I wish MS could
have just played nice. Anyone know if IE 5.5 beta has come to it's senses?
(You can easily test on the attachment to this bug)
That brings us back to the issue I raised in my "2000-06-09 10:32" comment -
Mozilla incorrectly merges attributes from both namespaces into one. The
application of a particular semantic to a namespace (Rec-html4 in this case)
should not effect how document fragments in a different namespace (XHTML in
this case) are rendered. As my atachment demonstrates, you should be able to
style HTML:H1 and XHTML:H1 independently, being as they are declared to be in
*different* namespaces.
So fine, support both namespaces, but if you do, I would ask to try and do it
properly (this is a gray area). And I will be the second to admit that I am
not an expert on this - I have read all I could find on the subject as of my
first commenting on this bug, but at that time even XML-Dev seemed to be rather
muddled WRT namespaces. I am beginning to wonder if _anyone_ has all the
answers. I still don't really understand partitions, nor how semantics for
unprefixed attribute name collisions between namespaces are resolved :-)
Reporter | ||
Comment 33•25 years ago
|
||
In response to the comments that IE5 recognizes only REC-html40: Do you have a
test case to demonstrate this? My testing of IE5 just now shows (what was
also my previous understanding) that IE5 doesn't actually care about the
namespace. IE5 applies HTML semantics to any namespace *qualifier* that is
"html". (Testing in IE 5.00.2919.6307 on Win98.) Was your test using "html"
and "xhtml" as the qualifiers? Could you test that your version of IE is also
doing this (i.e., using the qualifier rather than the namespace). If so, then
this is a non-issue.
Comment 34•25 years ago
|
||
IE just uses the prefix? Ewwee! Why the hell would they do something like
that? It is so blatantly againsed the standard.
My only test is the attachment to this bug which I have been referring to:
http://bugzilla.mozilla.org/showattachment.cgi?attach_id=6705
It (coincidentally) uses html as the prefix to the Rec-html4 namespace, which
would explain it working.
Under IE 5.00.2920.0000 using Windows 2000, only the html namespaced/prefixed
h1 appears as a heading (large). Neither styling for the html or xhtml is
recognized and thus both are black. If I remember correctly, as far as styling
is concerned, I think IE just has a problem with me using the (proper) '|'
vertical bar rather than ':' in the style sheet.
Under Mozilla 2000052508, both are rendered as headings (large). Both headings
appear in the blue xhtml style (demonstrating the attributes being merged).
So I guess that means REC-html4 might as well dissapear then.
Reporter | ||
Comment 35•25 years ago
|
||
Just to clarify (for the record), I think the cross-browser way of doing things
is:
* use the correct (XHTML) namespace
* use "html" as the namespace qualifier
and I don't think there are any X-browser design issues here.
Comment 36•25 years ago
|
||
Yes, now I remember, IE hardcodes recognition of html: as the HTML namespace
prefix *regardless* of what you've declared, and yes, that's totally in
violation of the point of the XML namespace spec.
Here's the issue, then:
1) We support doing the right thing.
2) As long as you use the html: prefix, it will work cross-browser.
3) Therefore, we *could* remove REC-html40 support from the browser, and at some
point we should as a cleanup step.
4) *However*, we have tons of XUL files in the product that use REC-html40, and
it would be a poor use of developers' and QA staffers' and external testers'
time at this late stage of development to start an effort to fix that all. Every
cycle put into that effort would be a cycle taken away from fixing REAL html40,
css1, and dom1 bugs, and that would reduce the degree of standards compliance in
the final product.
So here's what we're going to do:
1) leave the support for REC-html40 in for RTM solely so we needn't change all
our XUL files, explicitly deprecating it
2) aggressively educate and evangelize developers to use the correct namespace
identifier in the content and XUL they create; dbaron will create and post to
mozill.org a TechNote explaining how and why to do this and warning that we
intend to *drop* support for REC-html40 in a future release
3) post-RTM, clean up our own XUL files in the product to use the correct
namespace (I'll open a separate FUTURE bug on this)
4) after doing (3), we *may* fix this bug and drop REC-html40 from a post-RTM
release (I'll make this bug dependent on the bug for (3)
Vidur confirms that dbaron and I understand the issues here correctly. He notes
that there's one related issue which is that the reverse mapping of prefix (e.g.
html:) to URL (e.g. http://www.w3.org/1999/xhtml) is currently returning the
wrong one (REC-html40), and we *should* fix that. dbaron and I agree; vidur will
open a separate bug report on that and we will nominate that nsbeta2.
Removing [NEED INFO] and my own nsbeta3 nomination. Recommending nsbeta2-.
Marking FUTURE.
Comment 37•25 years ago
|
||
Note related bug 42278, "REC-html40 incorrectly returned as a result of reverse
mapping XHTML identifier."
Comment 38•25 years ago
|
||
Putting on [nsbeta2-] radar. Not critical to beta2. Putting on nsbeta3 keyword
radar.
Keywords: nsbeta3
Whiteboard: [nsbeta2-]
Comment 39•25 years ago
|
||
Eric/Nisheeth/Vidur, if the REC-html40 was removed, would you accept the fix?
Is there some other technical reason preventing the removal of REC-html40 (other
than the lack human resources)? The reason I am asking is because I could help
by passing the tree under a query-n-replace Perl script that would replace all
occurrences of REC-html40 with the right namespace, and automatically check-in
the modified files. Fixing this now will save from the reverse-mapping problem
that has just surfaced and the later headaches of explaining to people not to
use REC-html40.
Comment 40•25 years ago
|
||
My $.02: I think REC-html40 should be removed. In the long run, this would really be the best thing to do. IMHO, running a global search and replace sounds like a good thing to do.
Comment 41•25 years ago
|
||
My $.02: I think REC-html40 should be removed. In the long run, this would really be the best thing to do. IMHO, running a global search and replace sounds like a good thing to do.
Comment 42•25 years ago
|
||
Nisheeth wrote:
> Ian, if you agree with this stance, please remove the nsbeta2 keyword.
I completely agree (I had not originally realised that we already supported
the correct namespace). Seems like things moved on without me though. ;-)
Note that if our XUL and CSS files use the incorrect namespace, this will IMHO
have a much stronger impact than we might imagine. It is a fact of life that
most web authors do not code by looking up the correct behaviour and writing-to-
spec, they code by copy-and-paste. If Roger can do the mass search-and-replace
safely, then I would very strongly recommend that we use his fix.
Roger: If you do this, could you attach a diff so that we can double check
that only files that _should_ be changed did in fact get changed?
[BTW, does it make sense for an nsbeta2-,nsbeta3 bug to be marked Future? ;-) ]
OS: Linux → All
Hardware: PC → All
Reporter | ||
Comment 43•25 years ago
|
||
Regarding the search & replace - I think it's a good idea. However, it's not
quite as trivial as doing the search & replace and checking it in - you should
coordinate with other people who work on XUL files, since they may have a lot of
changes in their trees, and not want to deal with too many merge conflicts
(although in theory a change like this shouldn't cause merge conflicts, so
maybe it's not so bad). This means you should probably post to a newsgroup
(mozilla-xpfe??) about it, and maybe email a few of the key XUL people
(hyatt@netscape.com, ben@netscape.com, etc.).
Comment 44•25 years ago
|
||
Unfortunately, I am having difficulties attaching the Perl script and the
list of all the files that will be changed. Bugzilla seems not to handle
attachments very well (this is not first time that I experience such
difficulties.) With the Perl script, someone with a fast machine could launch
the mass search-and-replace, clobber-build, and tell us what is happening.
I am currently doing a clobber-build on my non-luxurious system and it may
take a while.
Comment 45•25 years ago
|
||
Comment 46•25 years ago
|
||
Comment 47•25 years ago
|
||
Finally, bugzilla accepted the attachments.
Using the script is pretty straightforward and it is worth giving a go if you
have a fast machine. To try *without effecting any replacement*, simply set
$just_show_me = 1 and from the directory above mozilla (on one line):
replace.pl 'http://www.w3.org/TR/REC-html40' 'http://www.w3.org/1999/xhtml'
mozilla > output.txt
This will give a listing similar to attachment 10081 [details].
Setting $just_show_me = 0 will *actually effect the changes in your tree*.
(replace.pl -i will do the replacements in a case-insensitive manner if this
is necessary?)
I am currently rebuilding my tree on Win32, and for safety, someone should
also try this on a different platform.
Comment 48•25 years ago
|
||
Comment 49•25 years ago
|
||
My clobber-build is finished and things seem to alright (viewer / profileWizard
/ mozilla / mailnews / composer).
Mindful of the fact that it is a search-n-replace of REC-htlm40 with something
else, is there somewhere specific in the product that I should look at?
Comment 50•25 years ago
|
||
To be safe, could you please propose this operation on
netscape.public.mozilla.xpfe, cc:ing hyatt and ben to be sure they see it,
before actually doing it on the tree? It would be great indeed if we can get
this fixed, but I'm concerned about changing a lot of files in the tree during a
time when folks are doing large #s of checkin's to meet the FC cutoff. Thanks
for your interest in this!!!
Comment 51•25 years ago
|
||
Thanks for your openeness. Albeit simple, this change is massive and frankly,
if it wasn't automated, it would be unlikely to do it at this crucial branching
point.
I have sent an email to a number of newsgroups including n.p.m.xpfe and CCed to
hyatt and ben. After running the script, I have checked with the 'Find' function
to see if there were files containing 'REC-html40'. One show up, layout.xml,
because the case didn't match.
Now that I have the tree built with the replacements (as listed in attachment
10084 [details]), is there something specific that I should particularly look at?
Are there some files (perhaps .html) that should be left alone with their
REC-html40 unchanged?
Comment 52•25 years ago
|
||
There was a catch... When the REC-html40 is part of the URI (as in the various
DOCTYPE) In fact, this would have been detected long ago had I attached a diff
as Ian suggested. I was delaying doing a diff because bugzilla scares me with
the attachments :-) I have now corrected the Perl script to avoid the case where
REC-html40 is part of a URI. I will attach a diff and a list of the REC-html40
left unchanged in the tree, if bugzilla permits... With both listings it would
be easier to tell what is replaced and what is left in. After check-in I could
then go and update http://www.mozilla.org/projects/mathml/authoring.html
to let math people know that Mozilla is indeed fully XML/XHTML 1.0 compliant :-)
Reporter | ||
Comment 53•25 years ago
|
||
Another thing you may not want to change is the C++ code that recognizes
REC-html40 as a namespace. (You can change the calls to CreateElementNS, etc.,
that use it.) That code should probably be changed by hand, only once we're
sure that we don't need to support it for backwards compatibility with existing
skins, etc. (It should certainly not be done immediately after checking in the
XUL changes.) Changing it by hand will also involve removing whatever allows
the 2 namespaces to be recognized as 1.
Comment 54•25 years ago
|
||
Comment 55•25 years ago
|
||
Comment 56•25 years ago
|
||
Comment 57•25 years ago
|
||
Attahing doesn't work reliably... Here are the occurences that will be
unchanged (with line-wrapping due to the bugzilla form):
mozilla/expat/expat.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/expat/gplelect.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/htmlparser/src/nsParser.cpp
===================================================================
"http://www.w3.org/TR/REC-html40/strict.dtd">
"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0//EN\"
\"http://www.w3.org/TR/REC-html40/strict.dtd\">",
"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0//EN\"
\"http://www.w3.org/TR/REC-html40/strict.dtd\">",
"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 6.0//EN\"
\"http://www.w3.org/TR/REC-html40/strict.dtd\">",
mozilla/extensions/transformiix/source/examples/include.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/extensions/transformiix/source/examples/sort.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/extensions/transformiix/source/examples/test.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/extensions/transformiix/source/examples/functions.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/webshell/tests/viewer/samples/test_weight.html
===================================================================
"http://www.w3.org/TR/REC-html40/loose.dtd">
mozilla/extensions/transformiix/source/xml/printer/HTMLPrinter.cpp
===================================================================
out <<endl<< " \"http://www.w3.org/TR/REC-html40/loose.dtd\">"
<<endl;
Comment 58•25 years ago
|
||
OK David, following your line of thoughts, I will temporarily change
kXHTMLNameSpaceURI in nsNameSpaceManager.cpp to the old REC-html40 stuff.
Therefore, as per Vidur's comments on 2000-02-29 11:15, this massive swap
will basically mirror the old behavior in reverse... Because the remaining
changes will be located in that single file (nsNameSpaceManager.cpp), a simpler
clean-up could be done upon confirmation that all is fine.
Reporter | ||
Comment 59•25 years ago
|
||
It might be better to leave nsNameSpaceManager.cpp alone and let vidur deal with
it (it's already covered by bug 42278, nsbeta2+). He already has a patch there.
Or maybe it should all be checked in together...
Comment 60•25 years ago
|
||
With these replacements and the change that I mention for
nsNameSpaceManager.cpp, the fix for bug 42278 is not necessary anymore.
Comment 61•25 years ago
|
||
Vidur, could you please clarify? Should I go ahead?
The script is ready and it will be doing:
cd relevant_dir
cvs update [prompt]
cvs add a_set_of_files [prompt]
cvs commit a_set_of_files [prompt]
Each prompt will give the opportunity to continue/skip/abort, and actions
will be logged.
Comment 62•25 years ago
|
||
Checkins in progress. Let the clean up begins and let's see how it goes...
Comment 63•25 years ago
|
||
I just read the entire diff, and with the exception of the changes in code which
have been mentioned above, I saw no change that seemed to me to be wrong.
Thanks Roger!
Comment 64•25 years ago
|
||
Apparently all went well. Is the final namespace of XUL still the mouthful
http://www.mozilla.org/keymaster/gatekeeper/there.is.only.xul ?
BTW, I understand the idea behind the W3C's format of namespaces
([w3c][year][spec]) is to include the year when the spec became a
recommendation. Mozilla may well do the same, taking the year of the invention
of the feature for historical values. This way we could all remember that it was
the year when we were hacking the lizard!
Reporter | ||
Comment 65•25 years ago
|
||
There are a bunch of changes that need to be made to the doc tree too.
(REC-html40 shows up as a namespace in lots of documentation.)
Comment 66•25 years ago
|
||
Ah... The more light at the situation, the more it appears that it was really
time to clean-up once for all. Thanks for letting the change happens.
The danger of leaving these temporary names is that there could be so many
legacy code and documents that the names would become a "standard" of their own.
A case in point is the fact that correcting the typo in 'referer' became
impossible due to a similar situation. So if XUL/XBL has to be cleaned-up as
well, I can go through an iteration of the perl script once more.
When you do these kinds of changes to the tree, you will also need to check
stuff that is NOT in cvs.mozilla.org but in gila.mozilla.org as well. For
example, the viewer XML debug demos are there (they are also in the normal
mozilla tree). I just checked in the fixes to gila for the debug demos (they are
now identical to the files in the mozilla tree).
Perhaps we should consider if these duplicates could be removed...
Then there is also the documentation and samples to consider... Searching for
"REC-html40" at www.mozilla.org just gave 82 hits.
Comment 68•25 years ago
|
||
Indeed, all these should be cleaned-up before things spread out of *.mozilla.org
and get out of control.
Comment 69•25 years ago
|
||
I just cleaned-up the doc tree as well. At the next refresh of the
mozilla.org website, the updates should show up. I think we can start
talking about XUL/XBL also :-) Do you really want to be typing:
http://www.mozilla.org/keymaster/gatekeeper/there.is.only.{xul,xbl}
in each and every of your XUL/XBL files ?!?
I understand these long-winded names were meant to be temporary.
Unfortunately, they haven't been revisited and more and more documents
are been produced with them. I am having that feeling that if we miss
this opportunity, it is going to be very hard (impossible?) to change
these names later. So what do you people think of the much simpler:
http://www.mozilla.org/2000/xul
http://www.mozilla.org/2000/xbl
All I need is a `go ahead' and I will run and monitor the Perl script
as I did with the previous batches of replacements.
Comment 70•25 years ago
|
||
Cc:ing some XUL people @netscape, hyatt and ben, and the mozilla.org's tree
custodians, brendan and waterson, to see what they think.
Comment 71•25 years ago
|
||
IIRC, "http://www.mozilla.org/keymaster/gatekeeper/there.is.only.xul" was
intentional, it was never intended to be a placeholder.
I have no problem with the status quo. I admit even to being somewhat fond of
the current NS. David could tell you more.
(btw: the current XBL NS is actually http://www.mozilla.org/xbl)
Comment 72•25 years ago
|
||
I didn't know that it was intentional and definitive. There are things that
outsiders like us don't know :-) Although the xmlns' "Notes" section at
http://www.mozilla.org/xpfe/xulref/windowref.html seems to suggest otherwise.
Comment 73•25 years ago
|
||
ah, had not seen that doc. :-)
Yes, David was the one who came up with the current NS as part of the whole
Ghostbusters theme surrounding XUL. He will cry if anyone makes him change it
;-)
Comment 74•25 years ago
|
||
Wow, namespace discussions appear to be passionate in general (at W3C, here and
elsewhere)! If that's the general feeling, then we can leave hyatt enjoy his
fond name for a while -- before it gets called ECMAsomething!! Be prepared :-)
Assignee | ||
Comment 75•25 years ago
|
||
Setting milestone to M17... Doesnt make sense for a bug with so much activity
to remain marked future, right? :-)
Target Milestone: Future → M17
Comment 76•25 years ago
|
||
Seems the bug has served its original purpose and can be marked
as FIXED? This will clear the way for some other bugs that
depend on this.
Now the ball has bounced back to Eric/Nisheeth/Vidur to decide whether
to completely get rid of REC-html40 from nsNameSpaceManager.cpp ...
Bug 42280 points out however that remnants of REC-html40 may still
exist in demos and the commercial tree, see? I am now leaning toward getting
rid of this REC-html40 while we can, so that we can be done with it, and
have the peace of mind that it will not come back and haunt us again.
Comment 77•25 years ago
|
||
nsbeta3-/future. let's not change support for this suddenly before a release,
but rather warn developers in advance when we know they are ready.
Whiteboard: [nsbeta2-] → [nsbeta2-] [nsbeta3-]
Target Milestone: M17 → Future
Updated•24 years ago
|
QA Contact: chrisd → petersen
Keywords: rtm
Assignee | ||
Comment 78•24 years ago
|
||
Marking rtm-.
Whiteboard: [nsbeta2-] [nsbeta3-] → [nsbeta2-] [nsbeta3-][rtm-]
I just learned that Opera also uses the incorrect HTML namespace, and ONLY the
incorrect one (bug has been filed for the Opera folks). I don't think we should
disable support for that anytime soon. Some kind of warning perhaps?
An example at http://www.opera.com/opera5/specs.html suggests that Opera 5 does
support the right namespace URL. Supposing that Opera 5 indeed supports the right
namespace, I don't think Opera is a reason enough for supporting the wrong namespace
URL.
I just downloaded the free Opera 5.01 for Windows (NT) and it does not seem to
understand the correct NS, at least not in the testcase 12/18/00 11:26.
Updated•24 years ago
|
Whiteboard: [nsbeta2-] [nsbeta3-][rtm-]
Comment 83•24 years ago
|
||
XMLNS related. Changing QA contact to bsharma@netscape.com
QA Contact: petersen → bsharma
Comment 84•23 years ago
|
||
What's the status of Opera 6.0 beta?
Opera 5.0 Carbon beta 2 supports the right namespace URL.
Opera Software has used the right namespace URL in their own press release
(http://www.opera.com/pressreleases/xhtml/20011026.xml), which suggests that
they support the right namespace on other platforms, too.
Comment 87•23 years ago
|
||
Marking FIXED since jst has now discontinued this REC-html40 in his checkin for
bug 134278.
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
Updated•22 years ago
|
QA Contact: petersen → rakeshmishra
You need to log in
before you can comment on or make changes to this bug.
Description
•