289363 - "View MathML source" context menu option broken

Status: RESOLVED FIXED

(Core :: MathML, defect)

Description

[u81239]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b2) Gecko/20050402 Firefox/1.0+
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b2) Gecko/20050402 Firefox/1.0+

As the bug description says, the "View MathML source" context menu option is
broken. It shows an empty source window.

I’d actually rather just see the option disappear. I think it’s a ‘bloat’
feature. As nobody noticed it so far anyway it doesn’t seem like the option is
much used, and I can achieve exactly the same thing by selecting the formula and
choosing ‘view selection source’ anyway.


~Grauw

Reproducible: Always

Steps to Reproduce:
1. Open page with MathML
2. Right-click on math to show context menu
3. Choose ‘View MathML source’

Actual Results:  
Empty window

Expected Results:  
Math source code

Comment 1

[u81239]

Ugh, what about using Unicode on bugzilla >_<. Again:

I'd actually rather just see the option disappear. I think it's a 'bloat'
feature. As nobody noticed it so far anyway it doesn't seem like the option is
much used, and I can achieve exactly the same thing by selecting the formula and
choosing 'view selection source' anyway.

~Grauw

Comment 2

[rbs]

Don't delude yourself into thinking that it is bloat. It is actually a small
auxiliary JavaScript function in the code of "View Selection Source". And I
should know because I implemented both. VSS came about because of VMS.

VSS itself doesn't fly to the face either. But those who get to notice it
immediately like it. VMS is even rarer because you have to be on MathML... the
likelihood of that is quite minuscule. So it is not surprising. All of this is
on purpose -- to keep them away from the mainstream.

There are things that get broken in MathML in general, and don't get notice
until much latter because MathML isn't mainstream (as yet).

Comment 3

[u81239]

Be that as it may be, it is broken, so it should be fixed then :).

I was just suggesting that if maintaining the feature proves to be too
difficult, it might as well be taken out, as it isn’t really that important. As
I said, 'VSS' pretty much duplicates the functionality.

By the way, if it’s about pretty-printing, it might very well be a good idea to
add a ‘pretty print source’ toggle to the view source window, so that it can
apply to everything XML (e.g. XHTML) and not just MathML (that would be nice,
actually :)).


~Grauw

Comment 4

[rbs]

Confirming. The problem happens here:
http://lxr.mozilla.org/mozilla/source/toolkit/components/viewsource/content/viewPartialSource.js#377

Comment 5

[rbs]

Attachment: patch

This is regression from bug 57600. It isn't present in the Aviary branch (and
thus Firefox 1.0) because different patches were checked in:
- The trunk got attachment 161587 [details] [diff] [review]
- Aviary got attachment 160008 [details] [diff] [review]

Comparing carefully the two, you will see that the Open() function in
attachment 160008 [details] [diff] [review] preserves a fallback to "about:blank" when the URI is null.
While this provision was (likely inadvertently) removed from attachment 161587 [details] [diff] [review].


As View MathML Source is doing:
377   var doc = getBrowser().contentDocument;
378   doc.open("text/html", "replace");
379   doc.write(source);
380   doc.close();

Tracing the code in the debugger showed that it falls in the above situation
because the caller is XUL with the principal = system principal for which the
URI is null. So the patch simply restores a fallback to about:blank as before.

Comment 6

[Johnny Stenback (:jst)]

Comment on attachment 191175 [details] [diff] [review]
patch

r+sr=jst

Comment 7

[rbs]

Comment on attachment 191175 [details] [diff] [review]
patch

Asking a=1.8b4. This was an inadvertent regression from  bug 57600. It had
side-effect of breaking document.open() when called from XUL.

Comment 8

[rbs]

Checked-in.