[bug#28832] [PATCH 1/3] gnu: Add emacs-json-reformat.

[ludo@gnu.org (Ludovic Courtès)]

Hi Oleg,

Oleg Pykhalov <go.wigust@gmail.com> skribis:

> ludo@gnu.org (Ludovic Courtès) writes:

[...]

>>> Did you have a chance to investigate the test failures?  It’s not
>>> confidence-inspiring ;-), so it would be good to at least have a link to
>>> an upstream bug report.
>
> Yes I looked at those tests.  Here is a little review of them.
>
> json-reformat-test:string-to-string basically just calls
> a reimplemented in Emacs 25 json-encode-string function.
>
> json-reformat-test:json-reformat-region-occur-error emulates error and
> produce a message.  This message is differ from Emacs 24 in symbol '`'.
>
> So, I think those test fails are harmless.  I attach the new patch with
> fixing those failing tests.

Indeed.

> From d589de21acb02d0fba7aee3b48e5f42d7bd8957e Mon Sep 17 00:00:00 2001
> From: Oleg Pykhalov <go.wigust@gmail.com>
> Date: Tue, 12 Dec 2017 01:41:08 +0300
> Subject: [PATCH] gnu: Add emacs-json-reformat.
>
> * gnu/packages/patches/emacs-json-reformat-fix-tests.patch: New file.
> * gnu/local.mk (dist_patch_DATA): Add this.
> * gnu/packages/emacs.scm (emacs-json-reformat): New variable.

[...]

> diff --git a/gnu/packages/patches/emacs-json-reformat-fix-tests.patch b/gnu/packages/patches/emacs-json-reformat-fix-tests.patch
> new file mode 100644
> index 000000000..23a239582
> --- /dev/null
> +++ b/gnu/packages/patches/emacs-json-reformat-fix-tests.patch
> @@ -0,0 +1,28 @@
> +Copyright © 2017 Oleg Pykhalov <go.wigust@gmail.com>
> +
> +This patch fixes tests for Emacs 25.

Please mention <https://github.com/gongo/json-reformat/issues/33> here.

OK with these changes!

> >> Besides, I wonder: shouldn’t ‘emacs-build-system’ define the
> >> ‘EMACSLOADPATH’ env. var. so we don’t have to carry all these -L flags?
> >> Is there any downside?
> 
> As I see from the documentation¹ EMACSLOADPATH is a list of directories
> with *.el files in it.  If we will use it, then it will be almost the
> same carring bunch of directories in package recipes, will it?

If ‘emacs-build-system’ sets ‘EMACSLOADPATH’ automatically, then
individual package definitions won’t need those -L flags.  Dunno if
there are good reasons not to do so.  Maybe Alex has an opinion?

Thanks,
Ludo’.