[oe] [meta-browser] CFT: Chromium 62 in meta-browser

[Martin Jansa]

On Wed, Nov 15, 2017 at 06:08:49PM +0100, Raphael Kubo da Costa wrote:
> Thanks for testing!
> 
> Martin Jansa <martin.jansa at gmail.com> writes:
> > Thanks for very good changes.
> >
> > Yesterday I've tried to build this on "bitbake world" servers and the build
> > for qemuarm64 failed with:
> [...]
> > longer log in http://errors.yoctoproject.org/Errors/Details/158898/
> [...]
> > I know what linking needs *a lot* of memory, but this server has 219GB
> > available and "only" 128GB from it can be taken by tmpfs, so there should
> > be still plenty RAM for successful linkage.
> 
> It had been a while since I'd last done an aarch64 build. I launched one
> here and linking failed with relocation errors. I've pushed 4e5abf8
> ("chromium: Apply -g -> -g1 workaround for aarch64 as well"), which is
> the same workaround I've been applying for other architectures, and
> everything went fine.

With the latest version from you (62.0.3202.94) I also got relocation
errors instead of the out-of-memory one, see:

http://errors.yoctoproject.org/Errors/Build/48426/

I'm trying now with g1 as you suggested.

> 
> > Build for x86_64 (on similar server) was successful.
> >
> > Build for x86 + musl failed even before start, because there is missing
> > space in the append:
> > recipes-browser/chromium/chromium_62.0.3202.89.bb:DEPENDS_append_x86 =
> > "yasm-native"
> > recipes-browser/chromium/chromium_62.0.3202.89.bb:DEPENDS_append_x86-64 =
> > "yasm-native"
> >
> > and unlike x86_64 it doesn't have trailing space in previous DEPENDS,
> > because of:
> > recipes-browser/chromium/chromium_62.0.3202.89.bb:DEPENDS_append_libc-musl
> > = " libexecinfo"
> >
> > Causing:
> > ERROR: Nothing PROVIDES 'libexecinfoyasm-native' (but
> > /home/jenkins/oe/world/shr-core/meta-browser/recipes-browser/chromium/
> > chromium_62.0.3202.89.bb DEPENDS on or otherwise requires it). Close
> > matches:
> >   libxinerama-native
> >   texinfo-native
> >   libyaml-native
> 
> Thanks for fixing the yasm bit.
> 
> As for musl: like I said in my original email and on GitHub, the musl
> build is currently broken. Carrying tons of patches in the recipe is an
> uphill battle, so unless someone takes the time to work with upstream I
> don't think it makes much sense to try to keep the build working.

OK, fair enough, maybe Khem will again chime in and fix it for musl (in
the meantime we might blacklist it for musl that people don't wait for
the build to fail).

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: Digital signature
URL: <http://lists.openembedded.org/pipermail/openembedded-devel/attachments/20171115/2eb04737/attachment-0002.sig>