| Summary: | cairo 1.14.12 fails test | ||
|---|---|---|---|
| Product: | cairo | Reporter: | Michael Rice <michael.rice> |
| Component: | general | Assignee: | Chris Wilson <chris> |
| Status: | RESOLVED MOVED | QA Contact: | cairo-bugs mailing list <cairo-bugs> |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | unspecified | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Michael Rice
2017-12-07 05:32:20 UTC
The test/README file has a lot of information on getting the tests to pass. It is unlikely you will get them all to pass as it is very dependent on the environment it is run in. Usually the easiest option is to open test/index.html and do a visual check that the output looks reasonable.
> I built the same packages in the same order exactly the same way (using a build
> script) on RHEL 7.3 without problem, so I'm perfectly willing to be told that
> some hidden dependency on RHEL 6.5 is being insufficiently met.
This is a good start. It may be the dependencies on RHEL 6.5 are too old to be be able to successfully pass the test. That doesn't mean your compiled cairo won't work. It just means the test output won't match the reference images.
I test and update ref images on Debian Testing. The further you get away from that platform the lower your chance of success.
> I'm attempting to build freetype 2.8.1, which requires harfbuzz (1.7.2),
> which requires cairo (1.14.12), which requires pixman (0.34.0). Freetype
> is intended for a non-standard prefix, so I'm building everything from source.
As far as I can see, harfbuzz only uses cairo for some stuff in the utils directory, not the harfbuzz library. Cairo will never be used by FreeType. The dependency is in the other direction.
Thanks for your comments Adrian, I can understand tests not running on a platform that is different from the reference platform. I run 'make test' on everything that provides it as a sanity check before running 'make install'. This is standard operating procedure of my build scripts. The goal is to provide a reference prefix for our projects. The prefix will span several generations of RHEL. There is other software that runs on these systems that precludes updating to the latest release, or even latest this quadrennium. I see from the test/README is a lot of information about why one would run other tests or specific tests in a specific manner. However, it doesn't have (or I glossed over it) a useful suggestion for the minimum set of tests that a) are guaranteed to pass and b) validate that the cairo build is viable. I'm building a mostly unattended build, so comparing index.html isn't _quite_ what I want. I really want a set of tests that if anything fails we do not install. Is there a minimum set that you would recommend that would be universally applicable? -- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/cairo/cairo/issues/295. |
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.