[XeTeX] Embarrassing fontspec
stranak at ufal.ms.mff.cuni.cz
Wed Jan 4 16:14:14 CET 2006
I believe we have encountered the same limitation of XeTeX due to my
silly specification of fonts for table collumns.
Will and Ross helped me back than, Ross pointing out that:
> Thus the \font command needs to be done again, every time, and you
> the available \font registers.
See the thread "Defining fonts for table columns in XeLaTeX".
On 4.1.2006, at 15:24, Jonathan Kew wrote:
> On 4 Jan 2006, at 5:57 am, Will Robertson wrote:
>> In trying to complete fontspec 1.9 before the new year properly
>> begins, I've come across a very strange problem that has me totally
> OK, I can now confirm that it is NOT fontspec that is at fault; this
> is a XeTeX bug. It will typically show up at some point when the total
> number of fonts you've loaded during the job (whether or not you
> actually used them all to print) gets significantly past 1000 or so.
> Because fontspec does so much "speculative" loading of fonts, to
> examine styles and features, you reach this figure in your job, though
> you don't come anywhere near it if you just use \font to load the
> styles you're actually going to print.
> I will release an update shortly with this bug fixed (along with a
> couple of others). In the meantime, you should be able to work around
> it by generating your figures in smaller groups, so that you don't
> access quite as many fonts during a single run (and thus run out of
> mojo). I realize that's inconvenient in a case like yours, Will;
> Incidentally, I noticed while trying to track this down that fontspec
> often seems to use \font with the exact same font specification two or
> three times in succession. Is there scope for some optimization of the
> behavior here? I didn't dig any deeper to try and figure out what it's
> really doing, but from inside XeTeX, it looked like repeated identical
> XeTeX mailing list
> postmaster at tug.org
More information about the XeTeX