gettext: msgcomm vs. msgcat and the handling of empty records

[Karl Eichwalder]

Karl Eichwalder <address@hidden> writes:

> +- Consider to add the --omit-header switch to msgcat; than remove
> +  msgcomm since all it can do is covered by msgcat.

Oops, an oversight by me occurred and I'd say that's a 'msgcat' bug.
While counting records 'msgcat' silently ignores untranslated -- thus
some of my strategies to deal with compendia, proofreading jobs and some
statistical issues are broken; e.g., it isn't possible to remove all
translations given .po file ('msghack' offered --empty to empty the
contents of a po .file).

'msgcat' must use the first occurrence of a record even if empty or
untranslated, probably even if fuzzy or obsolete, too; users interested
in certain flavors only should use 'msgattrib' first to select a subset.

Thus, for now please keep 'msgcomm'.  Here are three tests, only the
first one succeeds at the moment.

msgcomm-24
Description: Binary data

msgcat-8
Description: Binary data

msgcat-9
Description: Binary data

-- 
address@hidden (work) / address@hidden (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)