unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
* Re: [Emacs-diffs] master 81204b2: Port xref-tests to master branch
       [not found] ` <E1ayjGI-00045W-L4@vcs.savannah.gnu.org>
@ 2016-05-06 17:15   ` Dmitry Gutov
  2016-05-07 20:26     ` Dmitry Gutov
  0 siblings, 1 reply; 4+ messages in thread
From: Dmitry Gutov @ 2016-05-06 17:15 UTC (permalink / raw)
  To: emacs-devel, Paul Eggert

Hi Paul,

On 05/06/2016 08:09 PM, Paul Eggert wrote:
>  (defvar xref-tests-data-dir
>    (expand-file-name "data/xref/"
> -                    (file-name-directory (or load-file-name (buffer-file-name)))))
> +                    (getenv "EMACS_TEST_DIRECTORY")))

Is this change necessary?

Up until now I've always been able to just open a test file, M-x 
eval-buffer, and run the tests in the current Emacs instance with M-x ert.

This change seems like it will break that workflow.



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Emacs-diffs] master 81204b2: Port xref-tests to master branch
  2016-05-06 17:15   ` [Emacs-diffs] master 81204b2: Port xref-tests to master branch Dmitry Gutov
@ 2016-05-07 20:26     ` Dmitry Gutov
  2016-05-07 20:32       ` Paul Eggert
  0 siblings, 1 reply; 4+ messages in thread
From: Dmitry Gutov @ 2016-05-07 20:26 UTC (permalink / raw)
  To: emacs-devel, Paul Eggert; +Cc: Glenn Morris, Phillip Lord

Does anyone have better suggestions than

(defvar xref-tests-data-dir
   (expand-file-name "../../data/xref/"
                     (file-name-directory (or load-file-name 
(buffer-file-name)))))

?

'cd test && make test-file-name' is handy, but it's not a full 
replacement: when running ERT interactively, you get niceties like being 
able to drop into the debugger right where a test fails.

On 05/06/2016 08:15 PM, Dmitry Gutov wrote:
> Hi Paul,
>
> On 05/06/2016 08:09 PM, Paul Eggert wrote:
>>  (defvar xref-tests-data-dir
>>    (expand-file-name "data/xref/"
>> -                    (file-name-directory (or load-file-name
>> (buffer-file-name)))))
>> +                    (getenv "EMACS_TEST_DIRECTORY")))
>
> Is this change necessary?
>
> Up until now I've always been able to just open a test file, M-x
> eval-buffer, and run the tests in the current Emacs instance with M-x ert.
>
> This change seems like it will break that workflow.
>




^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Emacs-diffs] master 81204b2: Port xref-tests to master branch
  2016-05-07 20:26     ` Dmitry Gutov
@ 2016-05-07 20:32       ` Paul Eggert
  2016-05-07 20:44         ` Dmitry Gutov
  0 siblings, 1 reply; 4+ messages in thread
From: Paul Eggert @ 2016-05-07 20:32 UTC (permalink / raw)
  To: Dmitry Gutov, emacs-devel; +Cc: Glenn Morris, Phillip Lord

Dmitry Gutov wrote:
> Does anyone have better suggestions than
> 
> (defvar xref-tests-data-dir
>   (expand-file-name "../../data/xref/"
>                     (file-name-directory (or load-file-name (buffer-file-name)))))


I don't know the test architecture well and don't have any better suggestion.
That being said, the master branch should use the same pattern for all the tests
that consult the data file. I adjusted the xref tests to use the pattern I saw
elsewhere; if the pattern is wrong, we should fix it everywhere. Extra points if
you can figure out how to fix it without breaking merges from the emacs-25
branch, which uses a slightly different architecture in this area.



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Emacs-diffs] master 81204b2: Port xref-tests to master branch
  2016-05-07 20:32       ` Paul Eggert
@ 2016-05-07 20:44         ` Dmitry Gutov
  0 siblings, 0 replies; 4+ messages in thread
From: Dmitry Gutov @ 2016-05-07 20:44 UTC (permalink / raw)
  To: Paul Eggert, emacs-devel; +Cc: Glenn Morris, Phillip Lord

On 05/07/2016 11:32 PM, Paul Eggert wrote:

> That being said, the master branch should use the same pattern for all the tests
> that consult the data file.

Preferably yes, but if some authors are fine with using ERT from command 
line only, they could keep the current pattern if it's somehow 
preferable. I just don't want to see interactive-friendly patterns 
stomped out.

On a related note, most test code uses test/data and (getenv 
"EMACS_TEST_DIRECTORY"), but package-tests.el use the previous pattern 
(which I prefer) and test/lisp/emacs-lisp/package-resources/ (which 
seems rather inconsistent).

> I adjusted the xref tests to use the pattern I saw
> elsewhere; if the pattern is wrong, we should fix it everywhere.

Yes, no blame here. I just hadn't noticed it right away.

> Extra points if
> you can figure out how to fix it without breaking merges from the emacs-25
> branch, which uses a slightly different architecture in this area.

I don't think changes to tests on emacs-25 are going to be a significant 
problem much more.



^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2016-05-07 20:44 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <20160506170958.15680.28097@vcs.savannah.gnu.org>
     [not found] ` <E1ayjGI-00045W-L4@vcs.savannah.gnu.org>
2016-05-06 17:15   ` [Emacs-diffs] master 81204b2: Port xref-tests to master branch Dmitry Gutov
2016-05-07 20:26     ` Dmitry Gutov
2016-05-07 20:32       ` Paul Eggert
2016-05-07 20:44         ` Dmitry Gutov

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs.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).