155122 - Printing MathML is broken

Status: RESOLVED WORKSFORME

(Core :: MathML, defect)

Description

[Roland Mainz]

Printing MathML seems to be broken in trunk builds (tested with
2002-06-29-08-trunk), there are at least two detects...

Comment 1

[Roland Mainz]

Attachment: Page 1 of http://www.mozilla.org/projects/mathml/start.xml

Defect 1:
The "det" block is not enclosed into with vertical lines.

(ignore the colors, that's a bug in sdtimage when it writes GIFs... ;-( )

Comment 2

[Roland Mainz]

Attachment: Page 2 of http://www.mozilla.org/projects/mathml/start.xml

Defect 2:
The text block below the MathML block is shifted to the right - beyond the page
border.

Comment 3

[Roland Mainz]

Builds from the 1.0.1-branch suffer from the same defects. BAD

;-(

Comment 4

[Roland Mainz]

2002-06-10-14-trunk works OK; 2002-06-17-21-trunk is broken... ;-(

Comment 5

[Roland Mainz]

2002-06-14-21-trunk suffers from defect [1] but not defect [2].

Comment 6

[rbs]

> 2002-06-10-14-trunk works OK; 2002-06-17-21-trunk is broken

So this is a regression, any idea as to the checkin that caused that?

Comment 7

[rods (gone)]

it would be really helpful if someone could track it down to what day.. (or
check in)

Comment 8

[Roland Mainz]

I found an interesting detail:
Setting "View/Use Style" to "Mathematica" fixes both issue [1] and issue [2]
(tested with 2002-07-12-08-trunk on Solaris2.7/SPARC + attachment 91153 [details] [diff] [review] from bug
153080, but I assume that patch does not matter here).

Setting "View/Use Style" to "TeX" or "Basic Page Style" will result in both
issue [1] and issue [2] ...

Comment 9

[Roland Mainz]

Bonsai query for checkins into mozilla/layout between 06/14/2002 and 06/17/2002: 
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=MozillaTinderboxAll&branch=HEAD&branchtype=match&dir=mozilla%2Flayout&file=&filetype=match&who=&whotype=match&sortby=Who&hours=2&date=explicit&mindate=06%2F14%2F2002+00%3A00%3A00&maxdate=06%2F17%2F2002+23%3A59%3A00&cvsroot=%2Fcvsroot

I guess one of these checkins may responsible for the problem...

Comment 10

[rbs]

Any chance that this has been fixed by bug 154741? From the list of checkins
that you provided above, this bug looks like an upshot from karnaze's
NS_FRAME_COMPLETE changes. 

Comment 11

[Roland Mainz]

rbs wrote:
> Any chance that this has been fixed by bug 154741?

Unfortunately both issues still exist in 2002-08-04-08-trunk.

I wonder why these issues only occur for the CSS styles "Basic page style" and
"TeX" and not the "Mathematica" CSS style.
Any ideas ?

Comment 12

[Roland Mainz]

Printing MathML in "Basic Page style" and "TeX" more is becoming more and more
horked. Page 1 of start.xml now doesn't show any MathML chars anymore and even 
http://www.mozilla.org/projects/mathml/demo/texvsmml.xml now suffers from
similar defects as start.xml except in "Mathematica" style mode.

Comment 13

[Roland Mainz]

rbs:
Can't we make the "Mathematica" style the default for now until we'll figure-out
what's causing these problems, please ?

Comment 14

[Roland Mainz]

CC:'ing [e-mail] per bz's suggestion.

The problem is still there and has even become worse (just tested with
2002-11-08-13-trunk).
Workaround is still to select the menu item "View/Use Style.../Mathematica" ...

Comment 15

[Roland Mainz]

smontagu/jkeiser:
Any suggestions/ideas what may go wrong here?

Comment 16

[David Schweiger]

Roland - can you retest this please? Printing MathML works for me in mozilla1.4
again.

Comment 17

[Vedran Miletic]

retargeting

Comment 18

[Roland Mainz]

Tested Mozilla 1.6b on Linux (buildid 2003121012) - the problem is back... ;-((

Comment 19

[Roland Mainz]

Mozilla1.4 seems to be working... so something between 1.4 and 1.6b was
broken... again... ;-(((

I wish we would archive nightly builds longer than a few days ... that would
help narrowing things down a little bit better...

Comment 20

[Andrew Schultz]

I'm not seeing any difference between 1.4 and current trunk, although it does
look pretty bad.  I have the CMPS and Mathematica fonts installed, but I doubt
Mozilla is using them (the page wasn't different with/without the fonts).  Is
there more useful instructions than bug 35236 to get the fonts working?

Comment 21

[Jungshik Shin]

Roland, what printing module was this bug filed against? Is it Xprint? I guess
so because PS and PS-FT2 don't work at all (GetBoundingMetrics() is not
implemented for both). 

Comment 22

[Roland Mainz]

Jungshik Shin wrote:
> Roland, what printing module was this bug filed against? Is it Xprint? I guess
> so because PS and PS-FT2 don't work at all (GetBoundingMetrics() is not
> implemented for both).

The tests were made using Xprint, but I guess Win32 should have the problems,
too.

The Xprint server side can't be source of the problem since the same Xprt server
instance prints MathML OK with Mozilla1.1 but prints completely broken with
Mozilla1.6b - and the gfx/src/xprint/ + gfx/src/xlib sources weren't touched
much between mozilla1.4 and mozilla1.6b... I guess the problem hides in layout/
or content/ ...

Comment 23

[Frédéric Wang (:fredw)]

Is this bug still valid? It seems a duplicate of bug 28088 BTW...

Comment 24

[Frédéric Wang (:fredw)]

I've just tested with a fresh trunk build. The defects described in comments 1 and 2 are no longer present.