[llvm-dev] Contributing Bazel BUILD files similar to gn

Sun Nov 1 14:42:24 PST 2020

Yes, I agree, Renato’s thread seems to be converging and seems like a productive way to frame this!

-Chris

> On Oct 31, 2020, at 1:45 PM, Geoffrey Martin-Noble <gcmn at google.com> wrote:
> 
> Thanks for weighing in Chris :-)
> 
> Did you see Renato's thread proposing a more general policy for these sorts of less-supported components that are closely tied to the core project? (http://lists.llvm.org/pipermail/llvm-dev/2020-October/146249.html <http://lists.llvm.org/pipermail/llvm-dev/2020-October/146249.html>). I think that is probably the more fruitful direction to take this as it seems to also help formalize some existing practices in the LLVM ecosystem. Reading the "pitch" process it sounds like it is targeted at cases where an RFC has not been successful at reaching consensus. That thread at least seems to be heading in a direction of consensus while working out the details, so it's not clear to me whether a pitch is necessarily required? Then again, it's a more far-reaching policy, so I certainly wouldn't want it to be made in a way where not everyone can see and participate. Also I think the end result (if agreed) should be a policy documented on the LLVM website. I know one of my frustrations when putting this RFC together was not knowing what the general policies for this sort of thing were and just poking through the codebase and list archive to get a general sense. Given these points, perhaps a pitch is advisable regardless?
> 
> On Sat, Oct 31, 2020 at 12:55 PM Chris Lattner <clattner at nondot.org <mailto:clattner at nondot.org>> wrote:
> I agree with Tom.  This seems like a natural time to use the decision making process, and I encourage you to do so.  Some random points I’ve taken away from the threads:
> 
> 1) I don’t think this discussion has anything to do with the technical merits of bazel over cmake.  There is empirically a community of people who would benefit from this, and there is no proposal to replace cmake with bazel.  This is your "We'll spend more time contributing to LLVM instead :-D” point.  I would de-emphasize the technical points, because you’re not actually evangelizing the technology here, you’re making a practical pitch.
> 
> Agreed. I was deliberately attempting to not get into Bazel's technical merits. Perhaps I muddied the water a bit by mentioning Bazel catching layering issues. I was trying to point out a way in which this would enable those of us who already care about Bazel to leverage it to give back to the community, which would be made more effective if this went through.
> 
> 2) I don’t think the comparison to GN is very important.  It is prior art, but doesn’t mean that it is necessarily correlated to this decision.  However, it doesn’t seem to me that GN has been a problem in practice for the community, so perhaps there is something to learn from that.  For example, the llvm/utils/GN/README.rst file clearly labels GN support as “experimental and best effort”.  GN also seems maintained, and even has a "LLVM GN Syncbot” that is doing stuff.
> 
> 3) I think there is an important question (independent of Bazel) of “how do we foster new things” and “how do we support things that are empirically important to the community”?  We don’t want to burden all contributors to help out a few of them, but keeping something “in tree” and marked experimental doesn’t seem like a burden.
> 
> 4) The big question is: Should this be in the mono-repo, a separate llvm incubator project, or none of these?  What should the terms of support be, etc?
> 
> I think Renato's RFC is doing a good job trying to answer these questions in a more general sense, which should help avoid us having to rehash all this again each time :-)
> 
> If you frame this carefully, I think we can make a decision quickly on this and move on.  From my personal perspective, it seems pretty obvious that we should take this, we just need to get the terms clear.
> Glad to hear you're supportive. Hopefully this exercise will provide an impetus to clear some things up in a broader sense rather than just for this particular case.
>  
> For example, if I check in a cmake change and break a bazel builder, I shouldn’t be on the hook to fix it, and arguably shouldn’t even get emailed about it.
>  
> Yes. I don't think it's even arguable. Only people who have specifically indicated an interest in supporting the Bazel build should get any kind of notification about it failing.
> -Chris
> 
> 
> 
> > On Oct 30, 2020, at 9:42 AM, Tom Stellard via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
> > 
> > On 10/28/20 7:18 PM, Geoffrey Martin-Noble via llvm-dev wrote:
> >> Hi all,
> >> tl;dr: We'd like to contribute Bazel BUILD files for LLVM and MLIR in a side-directory in the monorepo, similar to the gn build.
> > 
> > Hi Geoffrey,
> > 
> > I think you've received some good feedback on this thread, but It also doesn't look like continuing discussion on this thread is going to lead to a consensus.  If you are still interested in making this change, I think it would be best to escalate this proposal into a "proposal pitch" described here[1].
> > 
> > There is also a template here[2] you can use.
> > 
> > [1] https://github.com/llvm/llvm-www/blob/master/proposals/LP0001-LLVMDecisionMaking.md <https://github.com/llvm/llvm-www/blob/master/proposals/LP0001-LLVMDecisionMaking.md>
> > [2] https://github.com/llvm/llvm-www/blob/master/proposals/LP0000-Template.md <https://github.com/llvm/llvm-www/blob/master/proposals/LP0000-Template.md>
> > 
> > -Tom
> > 
> >> Some of us have been working on open-source Bazel BUILD files for the LLVM Project. You may have seen us hanging out in the #build-systems discord channel. As you may know, Google uses Bazel internally and has maintained a Bazel BUILD of LLVM for years. Especially with the introduction of MLIR, we've got more and more OSS projects with a Bazel BUILD depending on LLVM (e.g. IREE <https://github.com/google/iree <https://github.com/google/iree>>and TensorFlow <https://github.com/tensorflow/tensorflow <https://github.com/tensorflow/tensorflow>>). We're also not the only ones using Bazel: e.g. PlaidML also has a Bazel BUILD of LLVM that they've borrowed from TF <https://github.com/plaidml/plaidml/blob/master/vendor/llvm/llvm.BUILD <https://github.com/plaidml/plaidml/blob/master/vendor/llvm/llvm.BUILD>>. Each of these projects has to jump through some weird hoops to keep their version of the Bazel BUILD files in sync with the code, which requires some fragile combination of scripts and human intervention. Instead, we'd like to move general-purpose Bazel BUILD files into the LLVM Project monorepo. We expect to follow the model of the GN build where these will be maintained by interested contributors rather than expecting the general community to maintain them.
> >> To facilitate and test this we've been developing a standalone repository that just has the Bazel BUILD files. It symlinks together the directory trees on top of a submodule as we would need in the monorepo to to avoid in-tree BUILD files. The configuration is at https://github.com/google/llvm-bazel <https://github.com/google/llvm-bazel>. We now have those in a good place and think they would be useful upstream.
> >> # Details
> >> ## What
> >> Bazel BUILD files for the LLVM, MLIR, and Clang (PR out for review <https://github.com/google/llvm-bazel/pull/72 <https://github.com/google/llvm-bazel/pull/72>>) subprojects, potentially expanding to others, as needed. Basically everything currently at https://github.com/google/llvm-bazel <https://github.com/google/llvm-bazel>.
> >> ## Where
> >> In https://github.com/google/llvm-bazelthe <https://github.com/google/llvm-bazelthe> BUILD files live in a single directory tree matching the structure of the overall llvm-project directory. For users, @llvm-project is a single Bazel repository <https://docs.bazel.build/versions/master/build-ref.html#repositories <https://docs.bazel.build/versions/master/build-ref.html#repositories>>that includes both LLVM and MLIR subprojects. To maintain this structure, we would probably want to put a `bazel` directory in the monorepo's utils directory <https://github.com/llvm/llvm-project/tree/master/utils <https://github.com/llvm/llvm-project/tree/master/utils>>, which currently only contains a directory for arcanist. This is different from gn, which is under the LLVM subproject's utils directory <https://github.com/llvm/llvm-project/tree/master/llvm/utils/gn <https://github.com/llvm/llvm-project/tree/master/llvm/utils/gn>>. We could similarly put the Bazel BUILD files under llvm/utils/bazel but have them be for the entire llvm project (the subsets that are supported). This seems like an odd structure to me, but I know that the CMake build for LLVM also builds the other subprojects <https://github.com/llvm/llvm-project/blob/529ac33197f6/llvm/tools/CMakeLists.txt#L34-L41 <https://github.com/llvm/llvm-project/blob/529ac33197f6/llvm/tools/CMakeLists.txt#L34-L41>>, so maybe this would be preferable.
> >> Alternatively we could split each subproject into a separate Bazel repository and put the Bazel build files under each subproject. I think this fragments the configuration of the BUILD without much benefit.
> >> ## Configurations
> >> We currently have configurations for Linux GCC and Clang, MacOS GCC and Clang, and Windows MSVC. Support for other configurations can be added as-desired, but supporting all possible LLVM build configurations is not the goal.
> >> ## Support
> >> Support would be similar to the gn build. Contributors could optionally update the Bazel BUILD files as part of their patches, but would be under no obligation to do so.
> >> ## Preserving History
> >> I don't *think* the history of llvm-bazel is interesting enough to try to merge it into the monorepo and I was planning to submit this as a single patch, but please let me know if you disagree.
> >> ## Benefits to the community
> >>  *
> >>    Projects that depend on LLVM and use the Bazel build system can
> >>    avoid duplicating fragile effort. We'll spend more time contributing
> >>    to LLVM instead :-D
> >>  *
> >>    Bazel is stricter than CMake in many ways (e.g. it requires that
> >>    even header dependencies be declared) and can catch layering issues
> >>    very easily. There's even an optional layering_check feature we
> >>    could turn on if its use would benefit the community. (though
> >>    currently the existing problematic layering makes it a burden to
> >>    maintain on our own). Even without that additional check, as I've
> >>    been keeping the Bazel build green, I've found and fixed a number of
> >>    layering issues in the past couple weeks (e.g.
> >>    https://reviews.llvm.org/rGb49787df9a <https://reviews.llvm.org/rGb49787df9a>
> >>    <https://reviews.llvm.org/rGb49787df9a535f03761c340dca7ec3ec1155133d <https://reviews.llvm.org/rGb49787df9a535f03761c340dca7ec3ec1155133d>>and
> >>    https://reviews.llvm.org/rGc17ae2916c <https://reviews.llvm.org/rGc17ae2916c>
> >>    <https://reviews.llvm.org/rGc17ae2916ccf45a0c1717bd5f11598cc4fff342a <https://reviews.llvm.org/rGc17ae2916ccf45a0c1717bd5f11598cc4fff342a>>).
> >> Here's a patch <https://reviews.llvm.org/D90352 <https://reviews.llvm.org/D90352>>adding the Bazel build system. It's basically just `cp -r llvm-bazel/llvm-bazel llvm-project/utils/bazel`.
> >> _______________________________________________
> >> LLVM Developers mailing list
> >> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> >> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev <https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
> > 
> > _______________________________________________
> > LLVM Developers mailing list
> > llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> > https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev <https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20201101/197627fa/attachment-0001.html>