From: Jiang Xin <worldhello.net@gmail.com>
To: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
Cc: "Junio C Hamano" <gitster@pobox.com>,
"Git List" <git@vger.kernel.org>,
"Jiang Xin" <zhiyou.jx@alibaba-inc.com>,
"Alexander Shopov" <ash@kambanaria.org>,
"Jordi Mas" <jmas@softcatala.org>,
"Matthias Rüster" <matthias.ruester@gmail.com>,
"Jimmy Angelakos" <vyruss@hellug.gr>,
"Christopher Díaz" <christopher.diaz.riv@gmail.com>,
"Jean-Noël Avila" <jn.avila@free.fr>,
"Bagas Sanjaya" <bagasdotme@gmail.com>,
"Alessandro Menti" <alessandro.menti@alessandromenti.it>,
"Gwan-gyeong Mun" <elongbug@gmail.com>, Arusekk <arek_koz@o2.pl>,
"Daniel Santos" <dacs.git@brilhante.top>,
"Dimitriy Ryazantcev" <DJm00n@mail.ru>,
"Peter Krefting" <peter@softwolves.pp.se>,
"Emir SARI" <bitigchi@me.com>,
"Trần Ngọc Quân" <vnwildman@gmail.com>,
"Fangyi Zhou" <me@fangyi.io>, "Yi-Jyun Pan" <pan93412@gmail.com>
Subject: Re: [PATCH v2 1/9] Makefile: sort "po/git.pot" by file location
Date: Thu, 19 May 2022 20:41:37 +0800 [thread overview]
Message-ID: <CANYiYbHBPVRto2ijF_f5VdSCem59=hjZj3kb=zF30hxrCPoT+Q@mail.gmail.com> (raw)
In-Reply-To: <220519.86tu9l6fw4.gmgdl@evledraar.gmail.com>
On Thu, May 19, 2022 at 5:17 PM Ævar Arnfjörð Bjarmason
<avarab@gmail.com> wrote:
>
>
> On Thu, May 19 2022, Jiang Xin wrote:
>
> > From: Jiang Xin <zhiyou.jx@alibaba-inc.com>
> >
> > Before feeding xgettext with more C souce files which may be ignored
> > by various compiler conditions, add new option "--sort-by-file" to
> > xgettext program to create stable message template file "po/git.pot".
> >
> > With this update, the newly generated "po/git.pot" will has the same
> > entries while in a different order. We won't checkin the newly generated
> > "po/git.pot", because we will remove it from tree in a later commit.
> >
> > Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
> > Signed-off-by: Junio C Hamano <gitster@pobox.com>
> > ---
> > Makefile | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/Makefile b/Makefile
> > index f8bccfab5e..83e968e2a4 100644
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -2706,6 +2706,7 @@ XGETTEXT_FLAGS = \
> > --force-po \
> > --add-comments=TRANSLATORS: \
> > --msgid-bugs-address="Git Mailing List <git@vger.kernel.org>" \
> > + --sort-by-file \
> > --from-code=UTF-8
> > XGETTEXT_FLAGS_C = $(XGETTEXT_FLAGS) --language=C \
> > --keyword=_ --keyword=N_ --keyword="Q_:1,2"
>
> I'm not opposed to this change, but between this and 2/9 I'm still
> unsure what the aim is exactly, and if the results are desired.
We will change the order of input files to feed xgettext to create
"po/git.pot" in patch 2/9, and we may change the order of input files
in future version. With the option "--sort-by-file" we can get a
stable "po/git.pot" and better compression ratios for future versions
of "po/*.po*".
> In 2/9 you correctly point out that we have messages we've missed due to
> LOCALIZED_C being platform-specific.
>
> That should never happen (although your implementation in 2/9 may have
> some small issues, I'll reply there separately), i.e. we should always
> have "make pot" generate the same po/git.pot from the same commit
> whether you're on linux, mac os x etc.
l10n translators may work on different platforms or have different
compiler conditions, and their contributions may have different base
templates (po/git.pot). It is hard for the l10n coordinator to choose
a base template to check contributions from different l10n
contributors.
> But AFAICT we have a "stable" sort order now, it's in whatever order we
> feed the files to xgettext, which ultimately comes down to e.g. the list
> of $(LIB_OBJS) in the Makefile.
>
> I've been looking over the libintl documentation to see what exactly
> these sort options do, and how they differ from the default, and it's
> not really described.
>
> AFAICT the xgettext behavior we have now is that we'll process the files
> we have in order, and for those files extract the messages we have as we
> see them.
>
> One fringe benefit of that is that e.g. "make pot
> XGETTEXT_INCLUDE_TESTS=Y" (which I think I've only ever used, and
> probably only ~10 years ago) will get added at the end, but now it'll be
> added wherever t/t0200 sorts.
>
> Then because compose the builtin objs and lib objs by concatenation, but
> don't $(sort) them in the Makefile most of this change is due to us
> e.g. sorting builtin/* before parse-options.c or whatever.
>
> But oddly we also have cases like this:
>
> strbuf_addf(&header, print_file_item_data.modified_fmt,
> _("staged"), _("unstaged"), _("path"));
>
> Before this we'd list those in that order in the git.pot, but now
> because of --sort-by-file we'll list any messages on the same line in
> sorted msgid order, not in the orderd they appear in. Another example is
> e.g. this in builtin/blame.c:
>
> OPT_STRING_LIST(0, "ignore-rev", &ignore_rev_list, N_("rev"), N_("ignore <rev> when blaming"))
>
> Where before we'd list them in that order, but now it's with "ignore.."
> before "rev".
These side effects of changing the order of entries in the same line
of the same source file have little effect on the l10n translation.
> I think this change would be easier to follow & explain if you first
> made this change:
>
> diff --git a/Makefile b/Makefile
> index 61aadf3ce88..3726fe8064a 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -2715,10 +2715,9 @@ XGETTEXT_FLAGS_SH = $(XGETTEXT_FLAGS) --language=Shell \
> --keyword=gettextln --keyword=eval_gettextln
> XGETTEXT_FLAGS_PERL = $(XGETTEXT_FLAGS) --language=Perl \
> --keyword=__ --keyword=N__ --keyword="__n:1,2"
> -LOCALIZED_C = $(C_OBJ:o=c) $(LIB_H) $(GENERATED_H)
> -LOCALIZED_SH = $(SCRIPT_SH)
> -LOCALIZED_SH += git-sh-setup.sh
> -LOCALIZED_PERL = $(SCRIPT_PERL)
> +LOCALIZED_C = $(sort $(C_OBJ:o=c) $(LIB_H) $(GENERATED_H))
> +LOCALIZED_SH = $(sort $(SCRIPT_SH) git-sh-setup.sh)
> +LOCALIZED_PERL = $(sort $(SCRIPT_PERL))
>
> ifdef XGETTEXT_INCLUDE_TESTS
> LOCALIZED_C += t/t0200/test.c
I do not need this intermediate commit and submit new generated
"po/git.pot" file to observe changes of the "po/git.pot" file, because
I have a diff driver install as described here:
https://github.com/git-l10n/git-po-helper/tree/main/contrib/diff-dirver
So I can tell the newly generated "po/git.pot" will have the same
entries while in a different order. If I want to see changes on raw
files, I can use command:
git diff --no-textconv
> Which would sort things within C, SH and Perl files (but not among
> them). Then this change would AFAICT only:
>
> * Change that "within one line" sort order, as noted above
> * Sort across C/SH/Perl.
>
> I'm mostly "meh" on the result, but it's also because I genuinely don't
> get what the goal was. Is it because in 2/9 you'll end up using
> $(FOUND_C_SOURCES), which we derive from either "git ls-files" or
> "find", the latter of which has an unstable sort order?
The goal is to have a constant order of entries in "po/git.pot" and
"po/*.po", so we can save the space of our repository by better
compression ratio on files inside "po/".
--
Jiang Xin
next prev parent reply other threads:[~2022-05-19 12:42 UTC|newest]
Thread overview: 110+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-03 13:23 [PATCH 0/9] Incremental po/git.pot update and new l10n workflow Jiang Xin
2022-05-03 13:23 ` [PATCH 1/9] Makefile: sort "po/git.pot" by file location Jiang Xin
2022-05-03 13:23 ` [PATCH 2/9] Makefile: generate "po/git.pot" from stable LOCALIZED_C Jiang Xin
2022-05-03 13:23 ` [PATCH 3/9] Makefile: have "make pot" not "reset --hard" Jiang Xin
2022-05-03 13:23 ` [PATCH 4/9] i18n CI: stop allowing non-ASCII source messages in po/git.pot Jiang Xin
2022-05-03 13:23 ` [PATCH 5/9] po/git.pot: don't check in result of "make pot" Jiang Xin
2022-05-03 13:23 ` [PATCH 6/9] po/git.pot: remove this now generated file, see preceding commit Jiang Xin
2022-05-03 13:23 ` [PATCH 7/9] Makefile: add "po-update" rule to update po/XX.po Jiang Xin
2022-05-03 13:23 ` [PATCH 8/9] Makefile: add "po-init" rule to initialize po/XX.po Jiang Xin
2022-05-03 13:23 ` [PATCH 9/9] l10n: Document the new l10n workflow Jiang Xin
2022-05-03 14:07 ` [PATCH 0/9] Incremental po/git.pot update and " Peter Krefting
2022-05-04 12:41 ` Jiang Xin
2022-05-04 14:35 ` Junio C Hamano
2022-05-04 14:51 ` Daniel Santos
2022-05-05 0:20 ` Jiang Xin
2022-05-05 22:00 ` Daniel Santos
2022-05-05 22:49 ` Junio C Hamano
2022-05-06 0:50 ` Jiang Xin
2022-05-05 0:07 ` Jiang Xin
2022-05-04 17:58 ` Junio C Hamano
2022-05-19 8:15 ` [PATCH v2 " Jiang Xin
2022-05-19 10:28 ` Ævar Arnfjörð Bjarmason
2022-05-19 14:32 ` Jiang Xin
2022-05-19 14:41 ` Ævar Arnfjörð Bjarmason
2022-05-23 1:25 ` [PATCH v3 " Jiang Xin
2022-05-23 7:15 ` Ævar Arnfjörð Bjarmason
2022-05-23 8:12 ` Ævar Arnfjörð Bjarmason
2022-05-23 13:42 ` Jiang Xin
2022-05-23 14:38 ` Ævar Arnfjörð Bjarmason
2022-05-23 16:13 ` Jiang Xin
2022-05-23 8:26 ` Jiang Xin
2022-05-23 15:21 ` [PATCH v4 " Jiang Xin
2022-05-23 18:19 ` Junio C Hamano
2022-05-26 14:50 ` [PATCH v5 00/10] " Jiang Xin
2022-05-26 14:50 ` [PATCH v5 01/10] Makefile: sort source files before feeding to xgettext Jiang Xin
2022-05-26 14:50 ` [PATCH v5 02/10] Makefile: generate "po/git.pot" from stable LOCALIZED_C Jiang Xin
2022-05-26 14:50 ` [PATCH v5 03/10] Makefile: have "make pot" not "reset --hard" Jiang Xin
2022-05-26 14:50 ` [PATCH v5 04/10] i18n CI: stop allowing non-ASCII source messages in po/git.pot Jiang Xin
2022-05-26 14:50 ` [PATCH v5 05/10] Makefile: remove duplicate and unwanted files in FOUND_SOURCE_FILES Jiang Xin
2022-05-26 14:50 ` [PATCH v5 06/10] po/git.pot: this is now a generated file Jiang Xin
2022-05-26 17:32 ` Junio C Hamano
2022-05-26 14:50 ` [PATCH v5 07/10] po/git.pot: don't check in result of "make pot" Jiang Xin
2022-05-26 14:50 ` [PATCH v5 08/10] Makefile: add "po-update" rule to update po/XX.po Jiang Xin
2022-05-26 14:50 ` [PATCH v5 09/10] Makefile: add "po-init" rule to initialize po/XX.po Jiang Xin
2022-05-26 14:50 ` [PATCH v5 10/10] l10n: Document the new l10n workflow Jiang Xin
2022-05-23 15:21 ` [PATCH v4 1/9] Makefile: sort "po/git.pot" by file location Jiang Xin
2022-05-23 15:21 ` [PATCH v4 2/9] Makefile: generate "po/git.pot" from stable LOCALIZED_C Jiang Xin
2022-05-23 15:21 ` [PATCH v4 3/9] Makefile: have "make pot" not "reset --hard" Jiang Xin
2022-05-25 22:19 ` Junio C Hamano
2022-05-25 22:24 ` Junio C Hamano
2022-05-26 1:10 ` Jiang Xin
2022-05-26 2:15 ` [PATCH] Makefile: dedup git-ls-files output to prevent duplicate targets Jiang Xin
2022-05-26 4:02 ` Junio C Hamano
2022-05-26 6:06 ` Jiang Xin
2022-05-26 6:23 ` Junio C Hamano
2022-05-26 7:04 ` Jiang Xin
2022-05-26 10:00 ` Ævar Arnfjörð Bjarmason
2022-05-26 11:06 ` Jiang Xin
2022-05-26 17:18 ` Junio C Hamano
2022-05-26 18:25 ` Ævar Arnfjörð Bjarmason
2022-05-26 19:00 ` Junio C Hamano
2022-05-26 19:17 ` Ævar Arnfjörð Bjarmason
2022-05-23 15:21 ` [PATCH v4 4/9] i18n CI: stop allowing non-ASCII source messages in po/git.pot Jiang Xin
2022-05-23 15:21 ` [PATCH v4 5/9] po/git.pot: this is now a generated file Jiang Xin
2022-05-23 15:21 ` [PATCH v4 6/9] po/git.pot: don't check in result of "make pot" Jiang Xin
2022-05-23 15:21 ` [PATCH v4 7/9] Makefile: add "po-update" rule to update po/XX.po Jiang Xin
2022-05-23 15:21 ` [PATCH v4 8/9] Makefile: add "po-init" rule to initialize po/XX.po Jiang Xin
2022-05-23 15:21 ` [PATCH v4 9/9] l10n: Document the new l10n workflow Jiang Xin
2022-05-23 1:25 ` [PATCH v3 1/9] Makefile: sort "po/git.pot" by file location Jiang Xin
2022-05-23 8:05 ` Junio C Hamano
2022-05-23 8:50 ` Jiang Xin
2022-05-23 1:25 ` [PATCH v3 2/9] Makefile: generate "po/git.pot" from stable LOCALIZED_C Jiang Xin
2022-05-23 8:05 ` Junio C Hamano
2022-05-23 1:25 ` [PATCH v3 3/9] Makefile: have "make pot" not "reset --hard" Jiang Xin
2022-05-23 7:28 ` Ævar Arnfjörð Bjarmason
2022-05-23 15:00 ` Jiang Xin
2022-05-24 0:56 ` Jiang Xin
2022-05-23 8:15 ` Junio C Hamano
2022-05-23 9:37 ` Jiang Xin
2022-05-23 1:25 ` [PATCH v3 4/9] i18n CI: stop allowing non-ASCII source messages in po/git.pot Jiang Xin
2022-05-23 1:25 ` [PATCH v3 5/9] po/git.pot: this is now a generated file Jiang Xin
2022-05-23 1:25 ` [PATCH v3 6/9] po/git.pot: don't check in result of "make pot" Jiang Xin
2022-05-23 7:26 ` Ævar Arnfjörð Bjarmason
2022-05-23 8:30 ` Jiang Xin
2022-05-23 8:35 ` Jiang Xin
2022-05-23 9:28 ` Ævar Arnfjörð Bjarmason
2022-05-23 1:25 ` [PATCH v3 7/9] Makefile: add "po-update" rule to update po/XX.po Jiang Xin
2022-05-23 1:25 ` [PATCH v3 8/9] Makefile: add "po-init" rule to initialize po/XX.po Jiang Xin
2022-05-23 1:25 ` [PATCH v3 9/9] l10n: Document the new l10n workflow Jiang Xin
2022-05-19 8:15 ` [PATCH v2 1/9] Makefile: sort "po/git.pot" by file location Jiang Xin
2022-05-19 8:53 ` Ævar Arnfjörð Bjarmason
2022-05-19 12:41 ` Jiang Xin [this message]
2022-05-19 8:15 ` [PATCH v2 2/9] Makefile: generate "po/git.pot" from stable LOCALIZED_C Jiang Xin
2022-05-19 9:18 ` Ævar Arnfjörð Bjarmason
2022-05-19 12:48 ` Jiang Xin
2022-05-19 8:15 ` [PATCH v2 3/9] Makefile: have "make pot" not "reset --hard" Jiang Xin
2022-05-19 9:43 ` Ævar Arnfjörð Bjarmason
2022-05-19 13:19 ` Jiang Xin
2022-05-19 14:06 ` Ævar Arnfjörð Bjarmason
2022-05-19 8:15 ` [PATCH v2 4/9] i18n CI: stop allowing non-ASCII source messages in po/git.pot Jiang Xin
2022-05-19 10:02 ` Ævar Arnfjörð Bjarmason
2022-05-19 8:15 ` [PATCH v2 5/9] po/git.pot: this is now a generated file Jiang Xin
2022-05-19 8:15 ` [PATCH v2 6/9] po/git.pot: don't check in result of "make pot" Jiang Xin
2022-05-19 8:15 ` [PATCH v2 7/9] Makefile: add "po-update" rule to update po/XX.po Jiang Xin
2022-05-19 10:07 ` Ævar Arnfjörð Bjarmason
2022-05-19 8:15 ` [PATCH v2 8/9] Makefile: add "po-init" rule to initialize po/XX.po Jiang Xin
2022-05-19 10:22 ` Ævar Arnfjörð Bjarmason
2022-05-19 8:15 ` [PATCH v2 9/9] l10n: Document the new l10n workflow Jiang Xin
2022-05-19 17:18 ` Junio C Hamano
2022-05-21 15:06 ` Jiang Xin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CANYiYbHBPVRto2ijF_f5VdSCem59=hjZj3kb=zF30hxrCPoT+Q@mail.gmail.com' \
--to=worldhello.net@gmail.com \
--cc=DJm00n@mail.ru \
--cc=alessandro.menti@alessandromenti.it \
--cc=arek_koz@o2.pl \
--cc=ash@kambanaria.org \
--cc=avarab@gmail.com \
--cc=bagasdotme@gmail.com \
--cc=bitigchi@me.com \
--cc=christopher.diaz.riv@gmail.com \
--cc=dacs.git@brilhante.top \
--cc=elongbug@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=jmas@softcatala.org \
--cc=jn.avila@free.fr \
--cc=matthias.ruester@gmail.com \
--cc=me@fangyi.io \
--cc=pan93412@gmail.com \
--cc=peter@softwolves.pp.se \
--cc=vnwildman@gmail.com \
--cc=vyruss@hellug.gr \
--cc=zhiyou.jx@alibaba-inc.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).