Re: [PATCH] t/meson.build: fix up tests for 'seen' branch

[Eli Schwartz <eschwartz93@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1494 bytes --]

On 10/21/24 5:14 PM, Taylor Blau wrote:
> On Mon, Oct 21, 2024 at 08:04:16AM +0200, Patrick Steinhardt wrote:
>> For our integration-style tests I'd be okay with not listing the files
>> individually, such that we instead use e.g. prove(1) to run all tests
>> via a single test target. It would be a regression in functionality as
>> we now cannot easily run e.g. "meson test t0000*", but at least we would
>> not have to maintain the list of test scripts anymore.
> 
> I quite like the ability to run tNNNN-*.sh individually when spelled
> like that (i.e. including the test number, but not the full script
> name).
> 
> I find that for the areas that I have worked on most over the years that
> I have the relevant test numbers committed by heart, but often don't
> know the full script name. So it is nice to be able to refer to them by
> number only.
> 
> Am I reading it write that this suggests that Meson would lose that
> ability?


If "meson test" simply ran the prove(1) utility then it would still be
possible to do:

```
meson test --test-args "--args --to --prove"
```

This would not actually do what you want, since prove(1) doesn't
implement "t0000*" style globbing, does it?

So your reading is correct -- meson's own globbing support relies on
meson knowing the list of all test names, which means you do have to
maintain that list somehow (and have that list update when you do "git
pull" and new tests materialize).


-- 
Eli Schwartz

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 963 bytes --]