bug#43754: [PATCH] configure help message of --with-json

bug#43754: [PATCH] configure help message of --with-json

[Shohei Yoshida]

[-- Attachment #1: Type: text/plain, Size: 311 bytes --]

version: git master(40b81f847f12cf141467a02858d1ac09589cea2f)

configure --help says

```
./configure --help
...
--with-json             don't compile with native JSON support
```

This is the opposite meaning. --with-json means "compile with native
JSON support"

Regards

-- 
Shohei YOSHIDA(syohex@gmail.com)

[-- Attachment #2: 0001-Fix-with-json-message.patch --]
[-- Type: text/x-patch, Size: 990 bytes --]

From 0dbe41c8ef7c69d8d5167e21b5cca45ab6060e58 Mon Sep 17 00:00:00 2001
From: Shohei YOSHIDA <syohex@gmail.com>
Date: Fri, 2 Oct 2020 09:30:53 +0900
Subject: [PATCH] Fix --with-json message

---
 configure.ac | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index ae0c0d2a2e..f0c8e5210f 100644
--- a/configure.ac
+++ b/configure.ac
@@ -453,7 +453,7 @@ OPTION_DEFAULT_ON([cairo],[don't compile with Cairo drawing])
 OPTION_DEFAULT_ON([xml2],[don't compile with XML parsing support])
 OPTION_DEFAULT_OFF([imagemagick],[compile with ImageMagick image support])
 OPTION_DEFAULT_ON([native-image-api], [don't use native image APIs (GDI+ on Windows)])
-OPTION_DEFAULT_IFAVAILABLE([json], [don't compile with native JSON support])
+OPTION_DEFAULT_IFAVAILABLE([json], [compile with native JSON support])
 
 OPTION_DEFAULT_ON([xft],[don't use XFT for anti aliased fonts])
 OPTION_DEFAULT_ON([harfbuzz],[don't use HarfBuzz for text shaping])
-- 
2.28.0

bug#43754: [PATCH] configure help message of --with-json

[Lars Ingebrigtsen]

Shohei Yoshida <syohex@gmail.com> writes:

> This is the opposite meaning. --with-json means "compile with native
> JSON support"

Thanks for the patch; applied to Emacs 28.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#43754: [PATCH] configure help message of --with-json

[황병희]

[just quick comment]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Shohei Yoshida <syohex@gmail.com> writes:
>
>> This is the opposite meaning. --with-json means "compile with native
>> JSON support"
>
> Thanks for the patch; applied to Emacs 28.

Personally i did mark this PR [43754] as *IMPORTANT*. Because currently
i use JSON things very useful in _real_ work/day-job.

Sincerely, JSON fan Byung-Hee

-- 
^고맙습니다 _地平天成_ 감사합니다_^))//

bug#43754: [PATCH] configure help message of --with-json

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Fri, 02 Oct 2020 03:14:56 +0200
> Cc: 43754@debbugs.gnu.org
> 
> Shohei Yoshida <syohex@gmail.com> writes:
> 
> > This is the opposite meaning. --with-json means "compile with native
> > JSON support"
> 
> Thanks for the patch; applied to Emacs 28.

Any reason why we use OPTION_DEFAULT_IFAVAILABLE for json, and not
OPTION_DEFAULT_ON?  How is it different from libpng or liblcms2 or
libgnutls?

bug#43754: [PATCH] configure help message of --with-json

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> Any reason why we use OPTION_DEFAULT_IFAVAILABLE for json, and not
> OPTION_DEFAULT_ON?  How is it different from libpng or liblcms2 or
> libgnutls?

I wonder whether it's because of this:

case $with_gnutls,$HAVE_GNUTLS in
  no,* | ifavailable,* | *,yes) ;;
  *) MISSING="$MISSING gnutls"
     WITH_IFAVAILABLE="$WITH_IFAVAILABLE --with-gnutls=ifavailable";;
esac
case $with_json,$HAVE_JSON in
  no,* | ifavailable,* | *,yes) ;;
  *) MISSING="$MISSING json"
     WITH_IFAVAILABLE="$WITH_IFAVAILABLE --with-json=ifavailable";;
esac

But if I understand autoconf (and I do not), that shouldn't really
"spread" to the OPTION_DEFAULT_ stuff, should it?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#43754: [PATCH] configure help message of --with-json

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: syohex@gmail.com,  43754@debbugs.gnu.org
> Date: Fri, 02 Oct 2020 16:28:38 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Any reason why we use OPTION_DEFAULT_IFAVAILABLE for json, and not
> > OPTION_DEFAULT_ON?  How is it different from libpng or liblcms2 or
> > libgnutls?
> 
> I wonder whether it's because of this:
> 
> case $with_gnutls,$HAVE_GNUTLS in
>   no,* | ifavailable,* | *,yes) ;;
>   *) MISSING="$MISSING gnutls"
>      WITH_IFAVAILABLE="$WITH_IFAVAILABLE --with-gnutls=ifavailable";;
> esac
> case $with_json,$HAVE_JSON in
>   no,* | ifavailable,* | *,yes) ;;
>   *) MISSING="$MISSING json"
>      WITH_IFAVAILABLE="$WITH_IFAVAILABLE --with-json=ifavailable";;
> esac

The $MISSING part is for telling the users some libraries they might
want were not available.  I'm saying that libjansson should be treated
the same as, say, libgnutls.

> But if I understand autoconf (and I do not), that shouldn't really
> "spread" to the OPTION_DEFAULT_ stuff, should it?

"Spread" in the sense that it should affect whether we use DEFAULT_ON
or DEFAULT_IFAVAILABLE?  No, I don't think so.

bug#43754: [PATCH] configure help message of --with-json

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> The $MISSING part is for telling the users some libraries they might
> want were not available.  I'm saying that libjansson should be treated
> the same as, say, libgnutls.
>
>> But if I understand autoconf (and I do not), that shouldn't really
>> "spread" to the OPTION_DEFAULT_ stuff, should it?
>
> "Spread" in the sense that it should affect whether we use DEFAULT_ON
> or DEFAULT_IFAVAILABLE?  No, I don't think so.

Yes, I'm just wondering if whoever added the OPTION_DEFAULT_AVAILABLE
did it by mistake by looking at this stuff, or whether there was a
reason for it.

Oh, hey, whaddayaknow, it was a patch I applied a couple months ago:

commit 43091e6c5069797ba17d2c7429e0122d3a5337d9
Author:     Noam Postavsky <npostavs@gmail.com>
AuthorDate: Fri Aug 14 19:31:16 2020 +0200
Commit:     Lars Ingebrigtsen <larsi@gnus.org>
CommitDate: Fri Aug 14 19:31:16 2020 +0200

    Make configure say so if we have "--with-json" but no jansson support
    
    * configure.ac (OPTION_DEFAULT_IFAVAILABLE): New macro.  Use it to
    define the --with-json option.  Add with_json and HAVE_JSON to the
    'MISSING' checks (bug#39953).

bug#39953 has the rationale behind the change.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#43754: [PATCH] configure help message of --with-json

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: syohex@gmail.com,  43754@debbugs.gnu.org
> Date: Sat, 03 Oct 2020 19:49:33 +0200
> 
> Oh, hey, whaddayaknow, it was a patch I applied a couple months ago:
> 
> commit 43091e6c5069797ba17d2c7429e0122d3a5337d9
> Author:     Noam Postavsky <npostavs@gmail.com>
> AuthorDate: Fri Aug 14 19:31:16 2020 +0200
> Commit:     Lars Ingebrigtsen <larsi@gnus.org>
> CommitDate: Fri Aug 14 19:31:16 2020 +0200
> 
>     Make configure say so if we have "--with-json" but no jansson support
>     
>     * configure.ac (OPTION_DEFAULT_IFAVAILABLE): New macro.  Use it to
>     define the --with-json option.  Add with_json and HAVE_JSON to the
>     'MISSING' checks (bug#39953).
> 
> bug#39953 has the rationale behind the change.

What I said there notwithstanding, it looks like --with-json is the
odd one out nowadays, as we handle all the other optional libraries
differently.  So it seems to me that we should switch to DEFAULT_ON
for this library.

Obviously, it is not a catastrophe to leave it as it is now...