[gentoo-dev] Handling optional, expensive variants of test suite

[gentoo-dev] Handling optional, expensive variants of test suite

[Sam James]

[-- Attachment #1: Type: text/plain, Size: 643 bytes --]

Hi!

Some packages like libffi, gcc support extended, slower versions of
their testsuites. In the past, I've seen both USE="expensive-tests" (I
think) and USE="test-full" (used in a few places in-tree atm) for this.

I sort of hate both suggestions but I'm open to what people think is
best, with a view to then making it a global USE flag then? Thoughts?

Note that I _do_ think there's value in exposing these because some of
the configurations I use, and I know others are deploying Gentoo for,
are where they want to make use of as many opportunities as possible
to find bad runtime behaviour (kernel and toolchain patching).

thanks,
sam

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 377 bytes --]

Re: [gentoo-dev] Handling optional, expensive variants of test suite

[Pacho Ramos]

[-- Attachment #1: Type: text/plain, Size: 1123 bytes --]

El vie, 09-08-2024 a las 16:40 +0100, Sam James escribió:
> Hi!
> 
> Some packages like libffi, gcc support extended, slower versions of
> their testsuites. In the past, I've seen both USE="expensive-tests"
> (I
> think) and USE="test-full" (used in a few places in-tree atm) for
> this.
> 
> I sort of hate both suggestions but I'm open to what people think is
> best, with a view to then making it a global USE flag then? Thoughts?
> 
> Note that I _do_ think there's value in exposing these because some
> of
> the configurations I use, and I know others are deploying Gentoo for,
> are where they want to make use of as many opportunities as possible
> to find bad runtime behaviour (kernel and toolchain patching).
> 
> thanks,
> sam

Hi!

I would opt for "test-full" as it looks for descriptive to me.
"expensive" will have different meanings (expensive in memory, CPU,
disk usage?) that are also likely to change in the future (for example,
needing 4Gb of RAM to run is probably less "expensive" now than ten
years ago).

In any case, if you prefer other terms, no problem.

Regards

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [gentoo-dev] Handling optional, expensive variants of test suite

[Joonas Niilola]

[-- Attachment #1.1: Type: text/plain, Size: 591 bytes --]

Hey,

On 9.8.2024 18.40, Sam James wrote:
> Hi!
> 
> Some packages like libffi, gcc support extended, slower versions of
> their testsuites. In the past, I've seen both USE="expensive-tests" (I
> think) and USE="test-full" (used in a few places in-tree atm) for this.
> 
> I sort of hate both suggestions but I'm open to what people think is
> best, with a view to then making it a global USE flag then? Thoughts?
> 

Count NSS in that list too! I've made a patch locally that uses
"tests-full" use flag, so I guess I'd vote for "test-full" to stay
consistent.

-- juippis

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 618 bytes --]

Re: [gentoo-dev] Handling optional, expensive variants of test suite

[Andrey Grozin]

On Fri, 16 Aug 2024, Joonas Niilola wrote:
> On 9.8.2024 18.40, Sam James wrote:
>> Some packages like libffi, gcc support extended, slower versions of
>> their testsuites. In the past, I've seen both USE="expensive-tests" (I
>> think) and USE="test-full" (used in a few places in-tree atm) for this.
>>
>> I sort of hate both suggestions but I'm open to what people think is
>> best, with a view to then making it a global USE flag then? Thoughts?
> Count NSS in that list too! I've made a patch locally that uses
> "tests-full" use flag, so I guess I'd vote for "test-full" to stay
> consistent.
Some tests of sympy take very long. They were commented out in the ebuild. 
I'd like to have an easy way to run all tests.

Andrey