You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@arrow.apache.org by Bhaskar Mookerji <mo...@gmail.com> on 2018/03/12 23:06:27 UTC
Adding support for Bazel?
Hey ya'll -
I've been following Apache Arrow for a bit and am catching up by building
the project locally on my machine. I was wondering: would there be any
interest in having a Bazel-based build available for arrow-cpp and its
bindings?
I would be happy to put up a polished proof of concept as a Github PR and
test it on Travis, as well as help maintain it in the future. I'm totally
aware that things like build tools and discussions about them introduce
overhead for project maintainers and can distract from other priorities. If
there are any serious objections, I'd love to hear them or write up a more
detailed proposal.
Background: I'm an engineer in the SF Bay Area with a recreational and
professional interest in the scientific computing and machine learning. I'm
really interested in helping
Thanks,
Buro
Re: Adding support for Bazel?
Posted by Phillip Cloud <cp...@gmail.com>.
I'm in agreement with Wes here. Packaging is much higher priority right
now. When we have an MVP for that we can revisit the build system issue,
which is definitely worth exploring.
On Wed, Mar 28, 2018, 20:50 Wes McKinney <we...@gmail.com> wrote:
> > As far as Bazel goes, I can put up a patch and discuss some of the
> pros/cons in the context of stated issues around packaging. It likely won't
> help with the release reliability questions, which seem separate.
>
> Indeed, the issues we have had with packaging are not (AFAICT) related
> to the build system we are using. We don't have an automated system to
> do nightly testing and automated package builds for release votes /
> post-release publishing.
>
> On Wed, Mar 28, 2018 at 8:21 PM, Bhaskar Mookerji <mo...@gmail.com>
> wrote:
> > oh gosh, I just saw this reply too!
> >
> > I've just caught up with the thread around the build+release/packaging.
> As
> > far as Bazel goes, I can put up a patch and discuss some of the
> pros/cons in
> > the context of stated issues around packaging. It likely won't help with
> the
> > release reliability questions, which seem separate.
> >
> > (Enjoy your vacation, Wes)
> >
> > Thanks!
> > Buro
> >
> > On Thu, Mar 22, 2018 at 8:13 PM, Wes McKinney <we...@gmail.com>
> wrote:
> >>
> >> hi Buro,
> >>
> >> Sorry for the delay in writing back!
> >>
> >> Our build systems would benefit a great deal from this sort of
> >> attention. Each component has its own build system at the moment, but
> >> it would be useful to have a Bazel-based build, for C++ at least but
> >> perhaps more components in the future. We've invested a lot of energy
> >> in building the CMake build system to this point and handling
> >> automatically building the dependencies; if Bazel ends up being more
> >> future-proof or more maintainable, that sounds good to me. In any
> >> case, we'd be glad to have you involved with the project.
> >>
> >> best,
> >> Wes
> >>
> >> On Mon, Mar 12, 2018 at 7:15 PM, Bhaskar Mookerji <mo...@gmail.com>
> >> wrote:
> >> > Truncation error:
> >> >
> >> > *helping contribute back to the ecosystem. Thanks for all the hard
> work!
> >> >
> >> > On Mon, Mar 12, 2018 at 4:06 PM, Bhaskar Mookerji <mookerji@gmail.com
> >
> >> > wrote:
> >> >
> >> >> Hey ya'll -
> >> >>
> >> >> I've been following Apache Arrow for a bit and am catching up by
> >> >> building
> >> >> the project locally on my machine. I was wondering: would there be
> any
> >> >> interest in having a Bazel-based build available for arrow-cpp and
> its
> >> >> bindings?
> >> >>
> >> >> I would be happy to put up a polished proof of concept as a Github PR
> >> >> and
> >> >> test it on Travis, as well as help maintain it in the future. I'm
> >> >> totally
> >> >> aware that things like build tools and discussions about them
> introduce
> >> >> overhead for project maintainers and can distract from other
> >> >> priorities. If
> >> >> there are any serious objections, I'd love to hear them or write up a
> >> >> more
> >> >> detailed proposal.
> >> >>
> >> >> Background: I'm an engineer in the SF Bay Area with a recreational
> and
> >> >> professional interest in the scientific computing and machine
> learning.
> >> >> I'm
> >> >> really interested in helping
> >> >>
> >> >> Thanks,
> >> >> Buro
> >> >>
> >
> >
>
Re: Adding support for Bazel?
Posted by Wes McKinney <we...@gmail.com>.
> As far as Bazel goes, I can put up a patch and discuss some of the pros/cons in the context of stated issues around packaging. It likely won't help with the release reliability questions, which seem separate.
Indeed, the issues we have had with packaging are not (AFAICT) related
to the build system we are using. We don't have an automated system to
do nightly testing and automated package builds for release votes /
post-release publishing.
On Wed, Mar 28, 2018 at 8:21 PM, Bhaskar Mookerji <mo...@gmail.com> wrote:
> oh gosh, I just saw this reply too!
>
> I've just caught up with the thread around the build+release/packaging. As
> far as Bazel goes, I can put up a patch and discuss some of the pros/cons in
> the context of stated issues around packaging. It likely won't help with the
> release reliability questions, which seem separate.
>
> (Enjoy your vacation, Wes)
>
> Thanks!
> Buro
>
> On Thu, Mar 22, 2018 at 8:13 PM, Wes McKinney <we...@gmail.com> wrote:
>>
>> hi Buro,
>>
>> Sorry for the delay in writing back!
>>
>> Our build systems would benefit a great deal from this sort of
>> attention. Each component has its own build system at the moment, but
>> it would be useful to have a Bazel-based build, for C++ at least but
>> perhaps more components in the future. We've invested a lot of energy
>> in building the CMake build system to this point and handling
>> automatically building the dependencies; if Bazel ends up being more
>> future-proof or more maintainable, that sounds good to me. In any
>> case, we'd be glad to have you involved with the project.
>>
>> best,
>> Wes
>>
>> On Mon, Mar 12, 2018 at 7:15 PM, Bhaskar Mookerji <mo...@gmail.com>
>> wrote:
>> > Truncation error:
>> >
>> > *helping contribute back to the ecosystem. Thanks for all the hard work!
>> >
>> > On Mon, Mar 12, 2018 at 4:06 PM, Bhaskar Mookerji <mo...@gmail.com>
>> > wrote:
>> >
>> >> Hey ya'll -
>> >>
>> >> I've been following Apache Arrow for a bit and am catching up by
>> >> building
>> >> the project locally on my machine. I was wondering: would there be any
>> >> interest in having a Bazel-based build available for arrow-cpp and its
>> >> bindings?
>> >>
>> >> I would be happy to put up a polished proof of concept as a Github PR
>> >> and
>> >> test it on Travis, as well as help maintain it in the future. I'm
>> >> totally
>> >> aware that things like build tools and discussions about them introduce
>> >> overhead for project maintainers and can distract from other
>> >> priorities. If
>> >> there are any serious objections, I'd love to hear them or write up a
>> >> more
>> >> detailed proposal.
>> >>
>> >> Background: I'm an engineer in the SF Bay Area with a recreational and
>> >> professional interest in the scientific computing and machine learning.
>> >> I'm
>> >> really interested in helping
>> >>
>> >> Thanks,
>> >> Buro
>> >>
>
>
Re: Adding support for Bazel?
Posted by Bhaskar Mookerji <mo...@gmail.com>.
oh gosh, I just saw this reply too!
I've just caught up with the thread around the build+release/packaging. As
far as Bazel goes, I can put up a patch and discuss some of the pros/cons
in the context of stated issues around packaging. It likely won't help with
the release reliability questions, which seem separate.
(Enjoy your vacation, Wes)
Thanks!
Buro
On Thu, Mar 22, 2018 at 8:13 PM, Wes McKinney <we...@gmail.com> wrote:
> hi Buro,
>
> Sorry for the delay in writing back!
>
> Our build systems would benefit a great deal from this sort of
> attention. Each component has its own build system at the moment, but
> it would be useful to have a Bazel-based build, for C++ at least but
> perhaps more components in the future. We've invested a lot of energy
> in building the CMake build system to this point and handling
> automatically building the dependencies; if Bazel ends up being more
> future-proof or more maintainable, that sounds good to me. In any
> case, we'd be glad to have you involved with the project.
>
> best,
> Wes
>
> On Mon, Mar 12, 2018 at 7:15 PM, Bhaskar Mookerji <mo...@gmail.com>
> wrote:
> > Truncation error:
> >
> > *helping contribute back to the ecosystem. Thanks for all the hard work!
> >
> > On Mon, Mar 12, 2018 at 4:06 PM, Bhaskar Mookerji <mo...@gmail.com>
> > wrote:
> >
> >> Hey ya'll -
> >>
> >> I've been following Apache Arrow for a bit and am catching up by
> building
> >> the project locally on my machine. I was wondering: would there be any
> >> interest in having a Bazel-based build available for arrow-cpp and its
> >> bindings?
> >>
> >> I would be happy to put up a polished proof of concept as a Github PR
> and
> >> test it on Travis, as well as help maintain it in the future. I'm
> totally
> >> aware that things like build tools and discussions about them introduce
> >> overhead for project maintainers and can distract from other
> priorities. If
> >> there are any serious objections, I'd love to hear them or write up a
> more
> >> detailed proposal.
> >>
> >> Background: I'm an engineer in the SF Bay Area with a recreational and
> >> professional interest in the scientific computing and machine learning.
> I'm
> >> really interested in helping
> >>
> >> Thanks,
> >> Buro
> >>
>
Re: Adding support for Bazel?
Posted by Wes McKinney <we...@gmail.com>.
hi Buro,
Sorry for the delay in writing back!
Our build systems would benefit a great deal from this sort of
attention. Each component has its own build system at the moment, but
it would be useful to have a Bazel-based build, for C++ at least but
perhaps more components in the future. We've invested a lot of energy
in building the CMake build system to this point and handling
automatically building the dependencies; if Bazel ends up being more
future-proof or more maintainable, that sounds good to me. In any
case, we'd be glad to have you involved with the project.
best,
Wes
On Mon, Mar 12, 2018 at 7:15 PM, Bhaskar Mookerji <mo...@gmail.com> wrote:
> Truncation error:
>
> *helping contribute back to the ecosystem. Thanks for all the hard work!
>
> On Mon, Mar 12, 2018 at 4:06 PM, Bhaskar Mookerji <mo...@gmail.com>
> wrote:
>
>> Hey ya'll -
>>
>> I've been following Apache Arrow for a bit and am catching up by building
>> the project locally on my machine. I was wondering: would there be any
>> interest in having a Bazel-based build available for arrow-cpp and its
>> bindings?
>>
>> I would be happy to put up a polished proof of concept as a Github PR and
>> test it on Travis, as well as help maintain it in the future. I'm totally
>> aware that things like build tools and discussions about them introduce
>> overhead for project maintainers and can distract from other priorities. If
>> there are any serious objections, I'd love to hear them or write up a more
>> detailed proposal.
>>
>> Background: I'm an engineer in the SF Bay Area with a recreational and
>> professional interest in the scientific computing and machine learning. I'm
>> really interested in helping
>>
>> Thanks,
>> Buro
>>
Re: Adding support for Bazel?
Posted by Bhaskar Mookerji <mo...@gmail.com>.
Truncation error:
*helping contribute back to the ecosystem. Thanks for all the hard work!
On Mon, Mar 12, 2018 at 4:06 PM, Bhaskar Mookerji <mo...@gmail.com>
wrote:
> Hey ya'll -
>
> I've been following Apache Arrow for a bit and am catching up by building
> the project locally on my machine. I was wondering: would there be any
> interest in having a Bazel-based build available for arrow-cpp and its
> bindings?
>
> I would be happy to put up a polished proof of concept as a Github PR and
> test it on Travis, as well as help maintain it in the future. I'm totally
> aware that things like build tools and discussions about them introduce
> overhead for project maintainers and can distract from other priorities. If
> there are any serious objections, I'd love to hear them or write up a more
> detailed proposal.
>
> Background: I'm an engineer in the SF Bay Area with a recreational and
> professional interest in the scientific computing and machine learning. I'm
> really interested in helping
>
> Thanks,
> Buro
>