Re: [PATCH v4 2/7] i18n: rebase: mark strings for translation

(cc-ing Duy because of a mention of his nice GETTEXT_POISON tweak[*])
Hi,

Jiang Xin wrote:

Mark strings in git-rebase.sh for translation. Jonathan offers a help
for reorgnization of the resolvemsg variable in 'git-rebase.sh', since
there is a likely message in git-am.sh, I update it in this commit for
consistency. And so does to 't/t0201-gettext-fallbacks.sh'.

Ah.  Looks like I tweaked the comma usage and sentence structure a
little.  Sorry, force of habit --- I shouldn't have.

Some test scripts are affected by this update, and would fail if tested
with GETTEXT_POISON switch turned on. Using i18n-specific test
functions, such as test_i18ngrep, in the related test scripts will fix
these issues.

If we're going to keep the changes together, here's how I would phrase
the commit message:

	Mark messages in git-rebase.sh for translation.  While doing this
	it was noticed that the comma usage and sentence structure of the
	resolvemsg was not quite right, so correct that and its cousins in
	git-am.sh and t/t0201-gettext-fallbacks.sh at the same time.

	Some tests would start to fail with GETTEXT_POISON turned on after
	this update.  Use test_i18ncmp and test_i18ngrep where
	appropriate to mark strings that should only be checked in the C
	locale output to avoid such issues.

Signed-off-by: Jiang Xin <redacted>
Reviewed-by: Stefano Lattarini <redacted>
Reviewed-by: Jonathan Nieder <redacted>

I haven't tested or reviewed this patch in detail, so even though it
looks good, I'd prefer it not to have my Reviewed-by.  (See
Documentation/SubmittingPatches: '"Reviewed-by:", unlike the other
extra tags, can only be offered by the reviewer'.)  If you'd like to
credit my help, something like "With advice from Jonathan." would be
fine.

[...]

Could we add a comment so others do not have to wonder what
human-readable message prompts the test_i18ngrep here?  e.g

	# "Does not point to a valid commit: invalid-ref"
	#
	# NEEDSWORK: This "grep" is fine in real non-C locales, but
	# GETTEXT_POISON poisons the refname along with the enclosing
	# error message.
	test_i18ngrep invalid-ref err

In the long run we may be able to turn this back to a "grep" again,
since any reasonable translation will keep the $onto_name somewhere
in the message.  But changing it to test_i18ngrep for now is the
right thing to do, until something like Duy's more sophisticated
version of GETTEXT_POISON arrives. (<-- [*])

Thanks,
Jonathan