Am Donnerstag, 17. März 2016 um 08:14:11, schrieb Günter Milde <mi...@lyx.org> > commit d0ef6afac08c86df4eb092fab49ba4dc7619aacb > Author: Günter Milde <mi...@lyx.org> > Date: Thu Mar 17 08:13:45 2016 +0100 > > cmake autotests: label aastex tests as unreliable > > document class aastex.cls replaced by aastex6.cls > (in CTAN and TL 2015-update around march 2016) > New layout required, see #10027 > > diff --git a/development/autotests/unreliableTests > b/development/autotests/unreliableTests
...
> -Sublabel: 1wrong 2output
> +Sublabel: varying_versions
> +# ========================
> +# Test depending on TeX distribution, package versions or OS
> +
> +# document class aastex.cls replaced by aastex6.cls
> +# (in CTAN and TL 2015-update around march 2016)
> +# New layout required, see #10027
> +export/templates/aastex.*
> +export/examples/aas_sample.*
> +
> +
> +Sublabel: wrong_output
This is the same as 'Sublabel: wrong output', because anything not in [a-zA-Z]
is treated as a separator.
But the sequence of displayed labels is not defined.
Opposite to this, the statement 'Sublabel: 1wrong 2output' defines the sequence.
It is pure luck that the labels now are still in correct order.
Each label has assigned a 'depth' flag, defaulting to '0'. '1wrong' assigns the
depth '1'. Labels are sorted according to depth
and according to the sequence appearing while parsing the *Tests controlling
files.
I prefer some labels with specified depth, be it only as an example.
Kornel
signature.asc
Description: This is a digitally signed message part.
