Re: ja/i18n-common-messages

[Junio C Hamano <gitster@pobox.com>]

Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:

> On Thu, Feb 03 2022, Junio C Hamano wrote:
>
>> * ja/i18n-common-messages (2022-01-31) 5 commits
>>  - i18n: fix some misformated placeholders in command synopsis
>>  - i18n: remove from i18n strings that do not hold translatable parts
>>  - i18n: factorize "invalid value" messages
>>  - SQUASH???
>>  - i18n: factorize more 'incompatible options' messages
>>
>>  Unify more messages to help l10n.
>>
>>  Will merge to 'next' after squashing the fix-up in.
>>  source: <pull.1123.v4.git.1643666870.gitgitgadget@gmail.com>
>
> I had a comment on the API direction in parse-options.c, which I think
> should be done differently, but I also think it would be fine to just
> change it up later:
>
> https://lore.kernel.org/git/220201.86a6fa9tmr.gmgdl@evledraar.gmail.com/
>
> I replied to v2 instead of v4 due to some (now fixed) mail delays at the
> time, but that comment still applies to the latest version.

I do not think the change at the parse-options level would mix well
with what this topic wants to do.  Large parts of the code this
series touches will have to be rewritten once again.

It will open us to new complaints we would not hear with this series
from users who first say "git cmd -a -b -c", get stopped with "a and
b are incompatible", then say "git cmd -a -c", get stopped again
with "a and c are incompatible", and utter "well you could have told
me upfront that these three are not to be used together" in
frustration.

I do not mind waiting for a few days to see what Jean-Noël would
say, but my take on this is that between starting from the current
code base and from the state after applying this series, there will
not be much difference in amount of the effort necessary to extend
CMDMODE to mark and detect combinations of incompatible options at
parse-options level. So I am inclined to merge this series down.