[dev-context] regression testing: 3 tex errors

Taco Hoekwater taco at elvenkind.com
Tue Feb 6 09:23:12 CET 2007


Sanjoy Mahajan wrote:
> Here are the three compilation errors (first one from each file with an
> error).  The Unicode and ec-iwona errors were mentioned before, so I'm
> skipping them here.  svn says who contributed the test files.
> 
> Taco, do you know if these tests are outdated, or whether a behavior has
> changed but shouldn't have?
> 
> -Sanjoy
> 
> 
> ======================================================================
> [r2 | taco | 2006-03-31 10:15:34 -0500 (Fri, 31 Mar 2006)]
> [the remaining unicode tables]
> 
> ./tex/context/base/unic-031.tex/test-001.tex:21: Undefined control sequence.
> \@default at greekEpsilonvaria ->\varia 
>                                      \greekEpsilon 

This is an actual bug, I think. I assume that the characters in 
enco-agr.tex should use \greekvaria instead of just \varia. Thomas?

> ======================================================================
> [r15 | taco | 2006-05-18 02:53:51 -0400 (Thu, 18 May 2006)]
> [\startalign tests]
> 
> ./tex/context/base/core-mat.tex/test-001.tex:25: Extra alignment tab has been changed to \cr.
> <template> $}\endtemplate 
>                           
> \doalignNC #1->&
>                 #1
> l.25 ... + b_1 y \NC = c_1 \NC d_1 u + e_1 v \NC =
>                                                    f_1 \NR

That one was diagnosed (and solved) by Aditya already.
(test was wrong)

> ======================================================================
> [r2 | taco | 2006-03-31 10:15:34 -0500 (Fri, 31 Mar 2006)]
> [the remaining unicode tables]
> 
> ./tex/context/base/unic-002.tex/test-001.tex:21: Missing \endcsname inserted.
> <to be read again> 
>                    \numexpr 
> \univ2 ...stroke \or \Lbar \else \ifcase \numexpr 
>                                                   #1-83\relax \bhook \or \un...

This is probably a bug in the vector that is triggered by
\showunicodetable: \strippedcsname does not like the second
\numexpr. It would be better to write the unic-002 as one
large \ifcase instead of attempting to be space-efficient
by using offsets. Hans?

Best, Taco


More information about the dev-context mailing list