Message ID | 1465315732-48921-1-git-send-email-cardoe@cardoe.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Doug Goldstein writes ("[PATCH for-4.7] travis: drop broken LLVM repos"): > LLVM repos are currently down so drop them from being installed so we > can get some testing back. This is unfortunate churn to be having in our tree. I'm happy for this removal commit to go in, but isn't there a way to deal with this that doesn't involve updating a metadata file in xen.git ? (This kind of external repo dependency is one reason why osstest likes to fetch its inputs from a very small number of different places, with many files and trees copied to xenbits.) Ian.
On Tue, Jun 07, 2016 at 12:08:52PM -0400, Doug Goldstein wrote: > LLVM repos are currently down so drop them from being installed so we > can get some testing back. > --- > .travis.yml | 2 -- > 1 file changed, 2 deletions(-) > > diff --git a/.travis.yml b/.travis.yml > index d2e1bec..5bae60d 100644 > --- a/.travis.yml > +++ b/.travis.yml > @@ -40,7 +40,6 @@ addons: > apt: > sources: > - ubuntu-toolchain-r-test > - - llvm-toolchain-precise > packages: > - zlib1g-dev > - libncurses5-dev > @@ -70,7 +69,6 @@ addons: > - gcc-aarch64-linux-gnu > - gcc-5 > - g++-5 > - - clang-3.8 > - seabios > # we must set CXX manually instead of using 'language: cpp' due to > # travis-ci/travis-ci#3871 > -- > 2.7.4 (Apple Git-66) > Can we just wait until llvm sorts itself out? I don't think the breakage is permanent, right? Wei.
On 6/7/16 12:17 PM, Ian Jackson wrote: > Doug Goldstein writes ("[PATCH for-4.7] travis: drop broken LLVM repos"): >> LLVM repos are currently down so drop them from being installed so we >> can get some testing back. > > This is unfortunate churn to be having in our tree. I'm happy for > this removal commit to go in, but isn't there a way to deal with this > that doesn't involve updating a metadata file in xen.git ? > > (This kind of external repo dependency is one reason why osstest likes > to fetch its inputs from a very small number of different places, with > many files and trees copied to xenbits.) > > Ian. > You can see it two ways. The tests for that specific version are always available since they're part of the repo. But if external dependencies go away then those tests are broken.
On 6/7/16 12:20 PM, Wei Liu wrote: > On Tue, Jun 07, 2016 at 12:08:52PM -0400, Doug Goldstein wrote: >> LLVM repos are currently down so drop them from being installed so we >> can get some testing back. >> --- >> .travis.yml | 2 -- >> 1 file changed, 2 deletions(-) >> >> diff --git a/.travis.yml b/.travis.yml >> index d2e1bec..5bae60d 100644 >> --- a/.travis.yml >> +++ b/.travis.yml >> @@ -40,7 +40,6 @@ addons: >> apt: >> sources: >> - ubuntu-toolchain-r-test >> - - llvm-toolchain-precise >> packages: >> - zlib1g-dev >> - libncurses5-dev >> @@ -70,7 +69,6 @@ addons: >> - gcc-aarch64-linux-gnu >> - gcc-5 >> - g++-5 >> - - clang-3.8 >> - seabios >> # we must set CXX manually instead of using 'language: cpp' due to >> # travis-ci/travis-ci#3871 >> -- >> 2.7.4 (Apple Git-66) >> > > Can we just wait until llvm sorts itself out? > > I don't think the breakage is permanent, right? > > Wei. > Its not permanent but it is now unknown when they will return. I figured I'd toss this patch out there if the community would rather have some build tests that pass instead of no build tests that run.
On Tue, Jun 07, 2016 at 01:09:04PM -0400, Doug Goldstein wrote: > On 6/7/16 12:20 PM, Wei Liu wrote: > > On Tue, Jun 07, 2016 at 12:08:52PM -0400, Doug Goldstein wrote: > >> LLVM repos are currently down so drop them from being installed so we > >> can get some testing back. > >> --- > >> .travis.yml | 2 -- > >> 1 file changed, 2 deletions(-) > >> > >> diff --git a/.travis.yml b/.travis.yml > >> index d2e1bec..5bae60d 100644 > >> --- a/.travis.yml > >> +++ b/.travis.yml > >> @@ -40,7 +40,6 @@ addons: > >> apt: > >> sources: > >> - ubuntu-toolchain-r-test > >> - - llvm-toolchain-precise > >> packages: > >> - zlib1g-dev > >> - libncurses5-dev > >> @@ -70,7 +69,6 @@ addons: > >> - gcc-aarch64-linux-gnu > >> - gcc-5 > >> - g++-5 > >> - - clang-3.8 > >> - seabios > >> # we must set CXX manually instead of using 'language: cpp' due to > >> # travis-ci/travis-ci#3871 > >> -- > >> 2.7.4 (Apple Git-66) > >> > > > > Can we just wait until llvm sorts itself out? > > > > I don't think the breakage is permanent, right? > > > > Wei. > > > > Its not permanent but it is now unknown when they will return. I figured > I'd toss this patch out there if the community would rather have some > build tests that pass instead of no build tests that run. OK. I'm fine with taking this patch. Wei. > > -- > Doug Goldstein >
>>> On 07.06.16 at 19:34, <wei.liu2@citrix.com> wrote: > On Tue, Jun 07, 2016 at 01:09:04PM -0400, Doug Goldstein wrote: >> On 6/7/16 12:20 PM, Wei Liu wrote: >> > On Tue, Jun 07, 2016 at 12:08:52PM -0400, Doug Goldstein wrote: >> >> LLVM repos are currently down so drop them from being installed so we >> >> can get some testing back. >> >> --- >> >> .travis.yml | 2 -- >> >> 1 file changed, 2 deletions(-) >> >> >> >> diff --git a/.travis.yml b/.travis.yml >> >> index d2e1bec..5bae60d 100644 >> >> --- a/.travis.yml >> >> +++ b/.travis.yml >> >> @@ -40,7 +40,6 @@ addons: >> >> apt: >> >> sources: >> >> - ubuntu-toolchain-r-test >> >> - - llvm-toolchain-precise >> >> packages: >> >> - zlib1g-dev >> >> - libncurses5-dev >> >> @@ -70,7 +69,6 @@ addons: >> >> - gcc-aarch64-linux-gnu >> >> - gcc-5 >> >> - g++-5 >> >> - - clang-3.8 >> >> - seabios >> >> # we must set CXX manually instead of using 'language: cpp' due to >> >> # travis-ci/travis-ci#3871 >> >> -- >> >> 2.7.4 (Apple Git-66) >> >> >> > >> > Can we just wait until llvm sorts itself out? >> > >> > I don't think the breakage is permanent, right? >> > >> > Wei. >> > >> >> Its not permanent but it is now unknown when they will return. I figured >> I'd toss this patch out there if the community would rather have some >> build tests that pass instead of no build tests that run. > > OK. I'm fine with taking this patch. Question then is - do we want this on just 4.7, or also unstable? Jan
On Wed, 2016-06-08 at 03:28 -0600, Jan Beulich wrote: > >>> On 07.06.16 at 19:34, <wei.liu2@citrix.com> wrote: > > On Tue, Jun 07, 2016 at 01:09:04PM -0400, Doug Goldstein wrote: > > >> Its not permanent but it is now unknown when they will return. I > figured > >> I'd toss this patch out there if the community would rather have > some > >> build tests that pass instead of no build tests that run. > > > > OK. I'm fine with taking this patch. > > Question then is - do we want this on just 4.7, or also unstable? > Both, I think? So that, AFAIU the issue, people using travis will have both their 4.7 _and_ unstable build staring being tested again? Dario
On Wed, Jun 08, 2016 at 03:28:56AM -0600, Jan Beulich wrote: > >>> On 07.06.16 at 19:34, <wei.liu2@citrix.com> wrote: > > On Tue, Jun 07, 2016 at 01:09:04PM -0400, Doug Goldstein wrote: > >> On 6/7/16 12:20 PM, Wei Liu wrote: > >> > On Tue, Jun 07, 2016 at 12:08:52PM -0400, Doug Goldstein wrote: > >> >> LLVM repos are currently down so drop them from being installed so we > >> >> can get some testing back. > >> >> --- > >> >> .travis.yml | 2 -- > >> >> 1 file changed, 2 deletions(-) > >> >> > >> >> diff --git a/.travis.yml b/.travis.yml > >> >> index d2e1bec..5bae60d 100644 > >> >> --- a/.travis.yml > >> >> +++ b/.travis.yml > >> >> @@ -40,7 +40,6 @@ addons: > >> >> apt: > >> >> sources: > >> >> - ubuntu-toolchain-r-test > >> >> - - llvm-toolchain-precise > >> >> packages: > >> >> - zlib1g-dev > >> >> - libncurses5-dev > >> >> @@ -70,7 +69,6 @@ addons: > >> >> - gcc-aarch64-linux-gnu > >> >> - gcc-5 > >> >> - g++-5 > >> >> - - clang-3.8 > >> >> - seabios > >> >> # we must set CXX manually instead of using 'language: cpp' due to > >> >> # travis-ci/travis-ci#3871 > >> >> -- > >> >> 2.7.4 (Apple Git-66) > >> >> > >> > > >> > Can we just wait until llvm sorts itself out? > >> > > >> > I don't think the breakage is permanent, right? > >> > > >> > Wei. > >> > > >> > >> Its not permanent but it is now unknown when they will return. I figured > >> I'd toss this patch out there if the community would rather have some > >> build tests that pass instead of no build tests that run. > > > > OK. I'm fine with taking this patch. > > Question then is - do we want this on just 4.7, or also unstable? > Both IMO. Wei. > Jan >
>>> On 07.06.16 at 18:08, <cardoe@cardoe.com> wrote: > LLVM repos are currently down so drop them from being installed so we > can get some testing back. > --- I was about to commit this, but you didn't provide an S-o-b. Jan
On 6/7/16 12:08 PM, Doug Goldstein wrote: > LLVM repos are currently down so drop them from being installed so we > can get some testing back. Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
diff --git a/.travis.yml b/.travis.yml index d2e1bec..5bae60d 100644 --- a/.travis.yml +++ b/.travis.yml @@ -40,7 +40,6 @@ addons: apt: sources: - ubuntu-toolchain-r-test - - llvm-toolchain-precise packages: - zlib1g-dev - libncurses5-dev @@ -70,7 +69,6 @@ addons: - gcc-aarch64-linux-gnu - gcc-5 - g++-5 - - clang-3.8 - seabios # we must set CXX manually instead of using 'language: cpp' due to # travis-ci/travis-ci#3871