Derrick Stolee <stolee@xxxxxxxxx> writes:
> On 11/23/2020 5:41 PM, Junio C Hamano wrote:
>> Emily Shaffer <emilyshaffer@xxxxxxxxxx> writes:
>>
>>>> OPT_BIT('l', "list", &actions, N_("list all"), ACTION_LIST),
>>>> + OPT_BOOL(0, "fixed-value", &fixed_value, N_("use string equality when matching values")),
>>> I'm not sure how to feel about this phrasing. I wonder if it would be
>>> clearer to say something like 'treat 'value_regex' as a literal string
>>> instead'? Hmmm.
>>
>> Update the document and help text with s/value_regex/value_pattern/
>> and say "use value_pattern as a fixed string, not an extended regexp",
>> perhaps?
>
> If I go about changing all documentation and error messages to say
> "value_pattern" instead of "value_regex", should I also update the uses
> in the *.po translation files? Or, should I leave them unmodified to
> trigger manual intervention by the translators?
If you do, you do not have to worry. The i18n/l10n coordinator will
update the po/git.pot file when we near the code freeze using an
automated tool that extracts strings from the sources, and the
po/xx.po files for languages are updated from the updated po/git.pot
mechanically with another tool, reusing unmodified entries, adding
new ones, and marking near-hit ones to help avoid unnecessary work
by translators.
I earlier thought that the "when matching values" phrase you chose
in this round, without such a clean-up, would be OK in the context
of this topic (which is depended on by a bugfix topic), but after
seeing how we need to clarify the way '!' negation prefix works in
the documentation that has deeply ingrained assumption that value is
matched using ERE, it may be necessary to bite the bullet and do the
regex->pattern now.
Thanks.
