214381 - Incorrect characters in prettyprint.dtd

Status: RESOLVED FIXED

(Core :: XML, defect)

Description

[Vincent Béron]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; fr-FR; rv:1.4) Gecko/20030630
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; fr-FR; rv:1.4) Gecko/20030630

There are 3 characters which made their way to the beginning of prettyprint.dtd,
before the license block, which should never be there and will probably cause
problems when Mozilla will try to parse the file.
The three characters in question are an ï, an » and an ¿, just before the <!--
of the license block.

Reproducible: Always

Steps to Reproduce:
1.
2.
3.

Comment 1

[Vincent Béron]

Attachment: Patch to remove incorrect characters in prettyprint.dtd

This patch fixes the problem.

Comment 2

[Vincent Béron]

Comment on attachment 128825 [details] [diff] [review]
Patch to remove incorrect characters in prettyprint.dtd

Requesting review from hjtoi-bugzilla.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Those bytes are a UTF-8 BOM, presumably added by sicking's text editor.  I sort
of doubt that they will cause any problems, and may be needed for the UTF-8 text
(the name in the comment) to not cause issues...

But then again, I could be wrong -- intl is not my strong suit.

Comment 4

[Heikki Toivonen (remove -bugzilla when emailing directly)]

I think you should take them out. The XML parser will assume UTF-8/16 if there
is no XML declaration, so if there are characters in UTF-8 it should be fine.

Confirming, and will review in a bit.

Comment 5

[Robert Kaiser]

This bug effectively prevents localization, as our MozillaTranslator treats that
DTD as invalid because of those charaters and leaves the translated file empty,
so that prettyprint doesn't work at all in localized builds.

requesting blocking1.5b because of that.

Comment 6

[Simon Montagu :smontagu]

Comment on attachment 128825 [details] [diff] [review]
Patch to remove incorrect characters in prettyprint.dtd

r=smontagu

Comment 7

[Vincent Béron]

Requesting integration of attachment #128825 [details] [diff] [review] on head by hjtoi-bugzilla.

Comment 8

[Michael Wolf]

I opened prettyprint.dtd in 2 different Unicode editors: UniPad and and UniRed.
UniPad shows in the Character Information Toolbar that those invalid characers
are  of the Arabic Presentation Forms-B subset. But UniPad doesn't give any
other information where you could see that there are invalid characters.
UniRed shows the invalid characters and displays that they were of cp-1250.
But in fact, those characters are not allowed to be there.

Comment 9

[bugzilla.mozilla.org]

is this the reason that I do not get a tree view for xml files anymore?

Since I tried a nightly yesterday I do not get prettyprint XML view, niether in
Mozilla nor in Firebird :((

Comment 10

[Heikki Toivonen (remove -bugzilla when emailing directly)]

Checked in.