[LEDE-DEV] Travis broken?
Etienne Champetier
champetier.etienne at gmail.com
Sat Aug 19 19:53:47 PDT 2017
Hi Philip, Alexander,
2017-08-19 14:39 GMT-07:00 Alexander Couzens <lynxis at fe80.eu>:
>
> On Sat, 19 Aug 2017 13:29:54 -0600
> Philip Prindeville <philipp_subx at redfish-solutions.com> wrote:
>
> > Github is trying to CI test a change to PHP7 and it’s never
> > completing, even after 50 minutes of build time:
> >
> > https://travis-ci.org/openwrt/packages/builds/266325516?utm_source=github_status&utm_medium=notification
> >
> > Is this a known issue from the recent round of changes?
> If you package has a huge dependency tree, there isn't a fix, except
> move from travis away to a private instance or jenkins.
PHP really has a huge dependency tree (2 ssl libs, 3 databases, ...)
For now we can just ignore these failures and let the author &
maintainer do the compile tests
(on top of the manual review)
>
>
> > It seems that a LOT of packages are being rebuilt, and I don’t
> > understand why. Isn’t the point of the latest reworking of Travis to
> > build against the pre-builds of the other packages, rather than do a
> > “big bang” rebuild of ALL dependencies?
> I don't think this is possible. The problem lies in the
> packaging of LEDE. As long LEDE doesn't package the development files
> into packages, the build of e.g. php-mysql would require to build mysql,
> because header files are not included in the mysql package.
Exactly, without "devel" packages we need to build every dependencies
My vision of Travis is to give feedback to PR author as fast as
possible for all the common/recurring problems,
so even if it works only 99% of the time it's still a big win IMO
Regards
Etienne
>
>
> Best,
> lynxis
> --
> Alexander Couzens
>
> mail: lynxis at fe80.eu
> jabber: lynxis at fe80.eu
> mobile: +4915123277221
> gpg: 390D CF78 8BF9 AA50 4F8F F1E2 C29E 9DA6 A0DF 8604
>
> _______________________________________________
> Lede-dev mailing list
> Lede-dev at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/lede-dev
>
More information about the Lede-dev
mailing list