<div dir="ltr"><div dir="ltr">Hi Serge!<div><br></div><div>Thanks for a lot of the answers here :)</div></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><br>
The proposed infrastructure makes this process smoother and non intrusive to the<br>
llvm-project codebase: all development can be done in a separate git repo,<br>
integration is controlled through cmake flags, and integration to<br>
clang/opt/bugpoint is built-in.<br>
<br></blockquote><div><br></div><div>Sounds reasonably nice. </div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
> That being said, perhaps it is worth it? But I think we need to call out why we<br>
> want it. I would also have expected something to llvm-dev for a change of this<br>
> magnitude. I didn't see anyone from the pass manager hierarchy on the reviews<br>
> and the final reviewer wasn't someone who contributes to these areas typically.<br>
<br>
I've (obviously) mentioned this development on llvm-dev, see<br>
<br>
<a href="http://lists.llvm.org/pipermail/llvm-dev/2019-September/135326.html" rel="noreferrer" target="_blank">http://lists.llvm.org/pipermail/llvm-dev/2019-September/135326.html</a><br>
<br>
Is there anything I should have done? Probably reaching llvm-dev before<br>
commiting. Reaching the right reviewers has always been a challenge to me, I had<br>
hoped that the mail to llvm-dev would trigger some subscription :-)<br></blockquote><div><br></div><div>Hrm. Probably finding some different reviewers, but I can't fault your attempts here. Usually you can look at the last few people to make substantial work in an area and loop them in via git log. :)</div><div><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
> In addition, what's with the OSX failure? It's currently turned off and was<br>
> breaking the bots, but does it mean that you don't expect this machinery to<br>
> work on OSX? That seems like a severely limiting factor for the project.<br>
<br>
I've setup github actions to test many configurations before merging [0], but missed one of<br>
them. I'm currently working on fixing that part.<br></blockquote><div> </div><div>Nice! Glad for the work here.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
> ; CHECK-EP-VECTORIZER-START-NEXT: Running pass: NoOpFunctionPass<br>
> +; CHECK-EXT: Running pass: {{.*}}::Bye on foo<br>
> <br>
> Why is this running on every test of the pass manager? It should be an example<br>
> run in the examples directory and not on by default? Same for every other PM<br>
> test. This seems like a bug?<br>
<br>
It's not. When the examples are active and if the appropriate cmake flag is set<br>
(which is not the case by default), the pass is linked in statically, and is run<br>
in the default pipeline. The CHECK-EXT prefix is disabled otherwise. That's one<br>
of the configuration I did test :-)<br>
<br></blockquote><div><br></div><div>I really don't think this is ideal. The examples directory shouldn't affect tests being run or not or in what way. Can we back this part out and talk about it a bit more? I don't think we should need to do this to test the functionality.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
> Thanks!<br>
<br>
Thanks for giving me the opportunity to clarify some obscure points, hope it<br>
helps !<br><br></blockquote><div><br></div><div>It very much did. Thanks for explaining.</div><div><br></div><div>-eric </div></div></div>