| Summary: | Do not build tests unless required | ||
|---|---|---|---|
| Product: | Telepathy | Reporter: | Gilles Dartiguelongue <gilles.dartiguelongue> |
| Component: | tp-glib | Assignee: | Telepathy bugs list <telepathy-bugs> |
| Status: | RESOLVED WONTFIX | QA Contact: | Telepathy bugs list <telepathy-bugs> |
| Severity: | enhancement | ||
| Priority: | medium | CC: | freedesktop-bugs |
| Version: | unspecified | Keywords: | patch |
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: | 0001-Compile-test-programs-only-when-needed.patch | ||
|
Description
Gilles Dartiguelongue
2009-08-05 15:27:28 UTC
I'm aware of the existence of check_PROGRAMS, but I don't think it's a good idea. The tests are not very time-consuming to build, and compiling them ensures that they don't bit-rot, so I'm inclined to say WONTFIX. building checks wihtout running them is a QA issue for us so I'll keep the patch for gentoo. FTR, I do run checks each time I upgrade a package because I really want to make sure I don't introduce obvious bugs to users :) What is your rationale for building tests unconditionally being considered a QA issue? Following discussion with other Telepathy hackers, in the absence of a compelling reason, this is WONTFIX. We consider it to be more important that the tests always compile successfully than that a tiny amount of CPU/disk is used to do so; our rationale is that it can sometimes be necessary to skip or retry "make check" (e.g. despite developers' best efforts, async tests sometimes have race conditions), but that it is never acceptable for the tests to fail to compile at all. I realise Gentoo and other source-based distributions are in the unusual position of wanting to compile as little as possible; you can patch this out, if you must. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.