On Wed, May 10, 2006 at 07:43:33AM +0200, Wouter Verhelst wrote: > retitle 365853 segfault when fonts produce nonvalid data > thanks > > On Tue, May 09, 2006 at 11:32:42PM -0400, Eric Dorland wrote: > > * Wouter Verhelst ([EMAIL PROTECTED]) wrote: > > > On Mon, May 01, 2006 at 06:30:22PM -0400, Eric Dorland wrote: > > > > * Wouter Verhelst ([EMAIL PROTECTED]) wrote: > > > > > On Mon, May 01, 2006 at 01:28:51AM -0400, Eric Dorland wrote: > > > > > > Run firefox --verbose, and see if FIREFOX_DSP is set. > > > > > > > > > > Nope. And I can still reproduce. > > > > > > > > Ok, how about with MOZ_DISABLE_PANGO=1? > > > > > > That seems to fix it. > > > > There have been some reports of crashes related to certain fonts. Any > > non-standard fonts installed? > > Nothing that's not in a Debian package. > > However, now that you mention fonts... > > I did have a problem with one font, 'Monospace', in that it was 'empty'; > its letters were of zero width and, thus, non-readable. When using > X-Chat, I had to switch to a different font to see *anything* in the > chat window (the text would be printed by X-Chat, but due to the fact > that the font's letters were of zero width and non-printing, the chat > window was empty). This bug seems to have been fixed; and when I try to > reproduce the bug in Firefox now, I can no longer do so. I'm saying > 'zero width', because when I typed something, my cursor wouldn't move, > while verifying the logs shows that X-Chat would think it was printed. > > Since the 'Monospace' font is declared as one of Firefox' default fonts > (the one for monospace text, obviously...), I assume the two are > related. So, I think we can safely assume that the bug is in Firefox' > handling of fonts, where it assumes that font data is always > well-formed, and crashes if not. > > Retitling appropriately. This sounds related to #365882.
-- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
