[zorg] r253237 - Always build the stage 2 cleanly.

David Blaikie via llvm-commits llvm-commits at lists.llvm.org
Mon Nov 16 10:43:43 PST 2015


On Mon, Nov 16, 2015 at 10:30 AM, Galina Kistanova <gkistanova at gmail.com>
wrote:

> No, this test is not implemented, as far as I know.
>
> The incremental stage 2 is an issue. Checking if the stage 1 compiler
> remains the same is a nice to have feature. I'd like to get the issue
> resolved and then see what we can do to improve.
>
> Do we even want to get the 2-stages build ran for the changes not related
> to the compiler?
> Regression tests, probably...
> Anyway, I'd like to get some solid use cases when we actually need to run
> 2-stages builder for changes not related to the compiler. Otherwise we
> shall just rethink on what kind of changes to trigger the builds.
>

Sometimes it's hard-to-impossible to tell. For example: updating the
wording of a comment changes the source files, but doesn't change the
resulting (non-debug*) binary, so we could still save a stage 2 build even
though we had no way to know to avoid the stage 1 build.

But this is all totally "nice to have" stuff, not necessary by any means.

* unless the comment doesn't change the number of lines, in which case even
a debug build's output won't change


>
> Thanks
>
> Galina
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151116/f9df8f33/attachment.html>


More information about the llvm-commits mailing list