I tried to print the MathML torture test page, and the MathML components come out as mostly black blocks with some white dots resembling slightly what they're supposed to be. (Their text descriptions came out OK.) (I looked, but I couldn't find this bug in the db...)

Attempted print with build 2000021117.

GetBoundingMetrics() has to be ported to gfx/ps (PostScript), as well as other platforms.

Just atached a diff and added 'helpwanted'. The patch is untested. A proper implementation requires: * support of more fonts (the current list of gSubstituteFonts[] in nsAFMObject.cpp is very limited for MathML purposes). * ability to switch fonts. * use of the charset converter manager to retrieve encoding for Unicode strings.

Is this a Win98-only issue or should this bug cover all platforms (except Xprint module on Unix, it can print MathML without problems) ?

Changing to All/All so that it can be used as a tracker. Printing doesn't work in Linux (GTK) because it relies on GfxPS which is lagging behind other flavours of GFX. GfxPS doesn't implement GetBoundingMetrics, and its implementation of GetTextDimensions is still not complete.

Just tried this under Mozilla 1.1 on Linux Mandrake 8.2, (PostScript) MathML printing functions, but there are several problems. The worst is that <mi> and <mo> items don't line up, to the effect that equals signs appear half a line above the variables. There are also horizontal spacing problems. Greek characters don't print at all. In expressions with a complicated expression under a square root, the square root symbol isn't scaled, and is placed in a wrong position. Besides that, MathML printing suffers from the other problems of Postscript printing under Linux. CSS styles are completely lost. Variables are printed as non-italic.

I've just tested with a Linux trunk build and can't see the problems described in comment 0 and comment 7. Can some people test with other platforms and confirm that this bug can be closed? In general, I see only one issue with MathML printing related to the thickness of some bars. I've opened bug 546044 for this.

> In general, I see only one issue with MathML printing related to the thickness > of some bars. I've opened bug 546044 for this. Another serious issue when printing document is linebreaking (see bug 534962). Closing this bug, since the specific issues described in this very old bug are no longer present. I suppose it is best to open separate bugs for problems related to MathML printing.