Running 'make install' is important to catch some VPATH problems
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
.travis.yml | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/.travis.yml b/.travis.yml
index 0328fcb8f1..61f0e38d40 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -105,12 +105,12 @@ before_install:
- if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew update && brew upgrade && brew install rpcgen yajl; fi
before_script:
- - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS
+ - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS --prefix=$(pwd)/../vroot
script:
# Many unit tests still fail on macOS, and there are a bunch of issues with
# syntax-check as well, so skip those steps on that platform for now
- - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi
+ - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi && make -j3 install
after_failure:
- echo '============================================================================'
--
2.14.3
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On Fri, 2018-02-23 at 12:01 +0000, Daniel P. Berrangé wrote: > Running 'make install' is important to catch some VPATH problems > > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> > --- > .travis.yml | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/.travis.yml b/.travis.yml > index 0328fcb8f1..61f0e38d40 100644 > --- a/.travis.yml > +++ b/.travis.yml > @@ -105,12 +105,12 @@ before_install: > - if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew update && brew upgrade && brew install rpcgen yajl; fi > > before_script: > - - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS > + - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS --prefix=$(pwd)/../vroot It would be nicer if the hard-coded parameters appeared before the dynamic ones here, in case we later find out we need per-OS overrides. Also I don't like "vroot" very much, I would prefer if you used "install" instead. But that's just my preference, so feel free to disregard it. > script: > # Many unit tests still fail on macOS, and there are a bunch of issues with > # syntax-check as well, so skip those steps on that platform for now > - - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi > + - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi && make -j3 install The install step should be right after building, so that the part of the command that is executed on all operating systems and the one that only applies to a subset are still visually separated. For the second hunk and with the above addressed, Reviewed-by: Andrea Bolognani <abologna@redhat.com> The first hunk might end up being dropped due to the comments on patch 4/4; if not, and with the comments addressed, the R-b applies to it too. -- Andrea Bolognani / Red Hat / Virtualization -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On Tue, Feb 27, 2018 at 04:23:01PM +0100, Andrea Bolognani wrote: > On Fri, 2018-02-23 at 12:01 +0000, Daniel P. Berrangé wrote: > > Running 'make install' is important to catch some VPATH problems > > > > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> > > --- > > .travis.yml | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/.travis.yml b/.travis.yml > > index 0328fcb8f1..61f0e38d40 100644 > > --- a/.travis.yml > > +++ b/.travis.yml > > @@ -105,12 +105,12 @@ before_install: > > - if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew update && brew upgrade && brew install rpcgen yajl; fi > > > > before_script: > > - - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS > > + - mkdir build && cd build && ../autogen.sh $CONFIGURE_ARGS --prefix=$(pwd)/../vroot > > It would be nicer if the hard-coded parameters appeared before > the dynamic ones here, in case we later find out we need per-OS > overrides. > > Also I don't like "vroot" very much, I would prefer if you used > "install" instead. But that's just my preference, so feel free to > disregard it. > > > script: > > # Many unit tests still fail on macOS, and there are a bunch of issues with > > # syntax-check as well, so skip those steps on that platform for now > > - - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi > > + - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi && make -j3 install > > The install step should be right after building, so that the part > of the command that is executed on all operating systems and the > one that only applies to a subset are still visually separated. I put the 'install' bit last, because that is the least likely bit to fail, so preferrable to have 'syntax-check' / 'check' run first to fail earlier I think it would be clearer if we just put a custom script: entry under OS-X in the matrix and got rid of the conditional check. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :| -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On Tue, 2018-02-27 at 15:44 +0000, Daniel P. Berrangé wrote: > > > script: > > > # Many unit tests still fail on macOS, and there are a bunch of issues with > > > # syntax-check as well, so skip those steps on that platform for now > > > - - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi > > > + - make -j3 && if [ "$TRAVIS_OS_NAME" != "osx" ]; then make -j3 syntax-check && make -j3 check; fi && make -j3 install > > > > The install step should be right after building, so that the part > > of the command that is executed on all operating systems and the > > one that only applies to a subset are still visually separated. > > I put the 'install' bit last, because that is the least likely > bit to fail, so preferrable to have 'syntax-check' / 'check' run > first to fail earlier It's also the part that takes the least time, so if syntax-check or check are going to fail, running install before them won't delay the result by any significant amount of time - not to mention that the overall completion time is easily dominated by waiting in the queue for a macOS builder. > I think it would be clearer if we just put a custom script: entry > under OS-X in the matrix and got rid of the conditional check. I disagree: separating the matrix definition and the script definition results in a much more readable configuration with zero duplicated information. In fact, we used to have everything jumbled together, but we moved away from that setup. -- Andrea Bolognani / Red Hat / Virtualization -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
© 2016 - 2025 Red Hat, Inc.