[XeTeX] (ConTeXt) trouble debugging a "Missing number" error

Mojca Miklavec mojca.miklavec.lists at gmail.com
Tue Dec 4 22:45:27 CET 2012

Dear Lars,

If you are willing to ignore the "please consider switching to LuaTeX"
;) replies, I would recommend you to send this question to the NTG
ConTeXt mailing list instead. There are more people there who can help
you diagnose the problem and possibly release a fix in case of a
recently introduced bug in ConTeXt. The error you saw is strictly

Please specify exactly which version of ConTeXt you are using, but
it's suggested that you try TeX Live 2012 and/or the standalone
context distribution with
    rsync -av rsync://contextgarden.net/distribution/setup/first-setup.sh .
    export PATH=/path/to/distribution/files/texmf-linux/bin:$PATH
and that you send as minimal example as possible (a tex file) that
triggers the problem. It might also help if you could replace
definitions like
    \definefont[SerifXXL][name:GentiumBookBasic at 12pt]
    \definefont[SerifXXL][file:lmroman10-regular.otf at 12pt]
(provided that it still fails to work) to make sure that users will be
able to run your example out-of-the-box without installing the latest
Gentium fonts.


On Tue, Dec 4, 2012 at 8:32 PM, Lars Huttar wrote:
> Hi all,
> I'm working with some code that was developed by someone else.
> The document I'm typesetting is now giving the following error:
> ! Missing number, treated as zero.
> <to be read again>
>                    \otr:1:27
> \OTRSETdoifcellelse ...er #1:\number #2\endcsname
>                                                   \@EA
> \secondoftwoarguments...
> \@@ar@@1 ...ellelse {\mofcolumns }\columnlastcell
>                                                   {\global \advance
> \columnl...
> \redoloop ->\expandrecursecontent
>                                   \endofloop
> <to be read again>
>                    {
> <inserted text>
>                 28
> ...
> l.12 \startabblist
>                   {
> ?
> I haven't been able to figure out what's wrong or how to fix it.
> I had just made changes to refactor the definitions of fonts. (I
> confirmed that these changes really are a necessary cause of the error:
> when I back out just these changes, the error goes away.)
> >From the above error message I understand the error was detected during
> the expansion of \startabblist. Here's the definition of abblist:
> % Abbreviations
> \definestartstop
>   [abblist]
>   [before={\setupbackgrounds[text][background=verticalline]
> \starttextbackground[text]
> \setuptab[headstyle={\switchtobodyfont[SansB,\SansBsize]},width=20mm,location=left]
> \startcolumnset[abbcolumn] \startalignment[flushleft,nothyphenated]},
>    after={\stopalignment \stopcolumnset \stoptextbackground
> \setupbackgrounds[text][background=off]}]
> >From that, one might guess that the font SansB or the font dimension
> \SansBsize were the problem. But those definitions haven't changed
> (though I would like to confirm that empirically with a run-time trace
> message).
> Here are the font settings changes that did occur:
> E16fonts.tex:
> changed this block:
>   \definefont[SerifXLI][name:GentiumBookBasic-Italic at 10pt]
>   \definefont[SerifXXL][name:GentiumBookBasic at 12pt]
>   \definefont[SerifXXLB][name:GentiumBookBasic-Bold at 12pt]
> to use size dimensions defined elsewhere:
> \definefont[SerifXLI][name:GentiumBookBasic-Italic at \XLfontsize]
> \definefont[SerifXXL][name:GentiumBookBasic at \XXLfontsize]
> \definefont[SerifXXLB][name:GentiumBookBasic-Bold at \XXLfontsize]
> E16settings.tex: I added these definitions:
> \define\XLfontsize{14pt}
> \define\XXLfontsize{16pt}
> There are modes involved, so things are a bit more complicated, but I
> believe that is the only effective difference.
> I also created a log file, which I'll attach.
> There is also a version with \tracingmacros=1, but it's 600K, so I'll
> put it at
> http://www.huttar.net/tmp/country-report-country_id-15-tracing.log
> Thanks for taking a look. Please let me know what further information I
> need to provide.
> I can provide the .tex files themselves upon request.
> I would especially like to know better techniques for debugging -- how I
> can induce TeX to tell me more about what's going on. I realize that as
> a macro language, this is inherently difficult with TeX.
> Lars

More information about the XeTeX mailing list