Re: [PATCH 1/5] i18n: keep the last \n even when text is poisoned
From: Jonathan Nieder <hidden>
Date: 2016-06-15 22:53:15
Nguyễn Thái Ngọc Duy wrote:
quoted hunk ↗ jump to hunk
--- a/gettext.c +++ b/gettext.c@@ -24,6 +24,16 @@ int use_gettext_poison(void) poison_requested = getenv("GIT_GETTEXT_POISON") ? 1 : 0; return poison_requested; } + +const char *poison_text(const char *msgid) +{ + int len = strlen(msgid); + if (len && msgid[len-1] == '\n') + return "# GETTEXT POISON #\n"; + else + return "# GETTEXT POISON #";
I realize this was not the motivation behind the above patch, but if the translation of some message has to end with a newline for git to function correctly, would we consider that a bug? I am of two minds on that: - on one hand, translators tend to be trustworthy, reasonable folks - on the other hand, anything we can do to make the translation process less fussy seems like time well spent The latter wins out for me, so I would prefer not to have this patch so the test suite can detect important newlines that should be not be part of the translatable string. Just my two cents, Jonathan