230579 - regular text seem to be rendered in white instead of black; < > = are not visible as well

Status: RESOLVED FIXED

(Core :: XML, defect, P3)

Description

[Alexandre]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040110
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040110

In the page at http://perso.wanadoo.fr/alexandre.l/poetry/poetry.xsl the < >
should be visible, same things for the = and the text between tags, the buttons
for expending and reducing nodes are not visible either. Note: the invisible
text can be seen by selecting it.

Reproducible: Always

Steps to Reproduce:
1. Load http://perso.wanadoo.fr/alexandre.l/poetry/poetry.xsl
2. Watch
Actual Results:  
The XML file isn't correctly rendered

Expected Results:  
It should have rendered the regular text in black

Comment 1

[Christian :Biesinger (don't email me, ping me on IRC)]

<bz> biesi: see
http://lxr.mozilla.org/seamonkey/source/content/xml/document/src/nsXMLContentSink.cpp#211
<bz> biesi: and
http://lxr.mozilla.org/seamonkey/source/content/xml/document/src/nsXMLContentSink.cpp#1173
<bz> biesi: if we have a special root, the fact that we have elements that have
presentation is ignored
<bz> biesi: which is totally broken
<bz> biesi: eg scripts will execute
<bz> biesi: stylesheets will load

Comment 2

[Boris Zbarsky [:bzbarsky]]

*** Bug 230581 has been marked as a duplicate of this bug. ***

Comment 3

[Alex Hirzel]

I tried to download the file in question to create a testcase, and it appears
that this effect isn't reproducable when the document is viewed locally.

Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.6b) Gecko/20031208

Comment 4

[Boris Zbarsky [:bzbarsky]]

That's because the link to the stylesheet is a relative url.

To sicking; we need a better way to handle this crap.

Comment 5

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

I've known about this problem for a while but i'm not really sure what to do
about it. This will only happen when trying to view XSLT and XBL documents. I
can see 3 "solutions":

1. Remove the code for 'special roots', this will cause XSLT and XBL documents
   to not be prettyprinted

2. Turn off scripts and stylesheets (and maybe other things too) when we
   detect a XSLT or XBL document. However this borders on not following some
   specs. (CC-ing hixie to get his oppinion on this)

3. Leave things as they are now.

Comment 6

[Hixie (not reading bugmail)]

2 is fine by me. We've already decided that we're not following the specs for
these, so... This only matters for the specific cases of XSLT and XBL, right?

Comment 7

[Hixie (not reading bugmail)]

Mind you, making sure the pretty printer always sets colours and backgrounds
together would solve this specific case, I think.

Comment 8

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Yes, only XBL and XSLT documents are treated this way, if any other type of
document contains a script or a stylesheet it won't be prettyprinted at all.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Attachment: Maybe like this?

Comment 10

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Comment on attachment 145913 [details] [diff] [review]
Maybe like this?

r=me though i'm not very thrilled about this.

The prettyprint code is growing more and more into a hack. What might be nice
to do is to once we detect that a page has a special root tell the document to
load as data (rather then 'view') and then let the css/script loaders query
that value.

However I'm not sure that that is worth it at this point. So i'd say go ahead
and check this in as is.

Comment 11

[Boris Zbarsky [:bzbarsky]]

> The prettyprint code is growing more and more into a hack.

"Growing"?  ;)

Note that things would be simpler (or at least clearer) if we had a
mWillPrettyPrint boolean about (starting at true, set to false by PIs or a
factored element without a special root)....

Comment 12

[Peter Van der Beken [:peterv]]

Comment on attachment 145913 [details] [diff] [review]
Maybe like this?

Sigh. Someone please file a bug about simplifying the prettyprinting stuff in
the XML content sink (with the suggestions from comments 10 and 11).

Comment 13

[Boris Zbarsky [:bzbarsky]]

Bug 241113 filed.

Comment 14

[Boris Zbarsky [:bzbarsky]]

Fixed on 1.8a trunk.