llvm dev - Jan 2019 - Buildbot for minimum supported GCC version? (seeing local build failures)

Hi,

Do we have any build bots for the currently minimum required version of
GCC? I'm currently seeing a build failure locally using v4.8.4, due to a
compiler bug fixed in 4.9 (see GCC bug id 57824), coming from
OptionalTest.cpp, relating to raw string literals in macros not being able
to encompass multiple lines. The change in question (r351548) was
introduced 3 days ago, and as far as I can see hasn't flagged up a build
bot failure.

I'm also seeing a build failure coming from llvm-xray, although in this
case it looks like it is actually a change in a header somewhere that is
causing it. I haven't yet tracked down the revision in question for this
(it's possible it's r351701 (and not fixed by r351728) but also possible
that something on my machine has got messed up), but removing the contents
of xray-converter.cpp means I can at least build successfully.

Regards,

James
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20190121/d9110407/attachment.html>

Hello James,

I believe r351756 should fix the build error in OptionalTest that you
are seeing.

As for the buildbots, I agree we should have a buildbot building with
the minimum supported compiler version. If noone cares about that
configuration enough to put one up, maybe we should consider dropping
support for that version?

pavel

On Mon, 21 Jan 2019 at 03:51, James Henderson
<jh7370.2008 at my.bristol.ac.uk> wrote:
>
> Hi,
>
> Do we have any build bots for the currently minimum required version of
GCC? I'm currently seeing a build failure locally using v4.8.4, due to a
compiler bug fixed in 4.9 (see GCC bug id 57824), coming from OptionalTest.cpp,
relating to raw string literals in macros not being able to encompass multiple
lines. The change in question (r351548) was introduced 3 days ago, and as far as
I can see hasn't flagged up a build bot failure.
>
> I'm also seeing a build failure coming from llvm-xray, although in this
case it looks like it is actually a change in a header somewhere that is causing
it. I haven't yet tracked down the revision in question for this (it's
possible it's r351701 (and not fixed by r351728) but also possible that
something on my machine has got messed up), but removing the contents of
xray-converter.cpp means I can at least build successfully.
>
> Regards,
>
> James

Based on other discussions re. C++14 support etc, I think we will be
dropping support for this version of GCC at some point in the future, but
not yet! As a point of principle though, I think whatever becomes our
minimum supported version should have a buildbot for it (same goes for
clang and MSVC, although I know we have an MSVC one already).

James

On Mon, 21 Jan 2019 at 18:26, Pavel Labath <labath at google.com> wrote:
> Hello James,
>
> I believe r351756 should fix the build error in OptionalTest that you
> are seeing.
>
> As for the buildbots, I agree we should have a buildbot building with
> the minimum supported compiler version. If noone cares about that
> configuration enough to put one up, maybe we should consider dropping
> support for that version?
>
> pavel
>
> On Mon, 21 Jan 2019 at 03:51, James Henderson
> <jh7370.2008 at my.bristol.ac.uk> wrote:
> >
> > Hi,
> >
> > Do we have any build bots for the currently minimum required version
of
> GCC? I'm currently seeing a build failure locally using v4.8.4, due to
a
> compiler bug fixed in 4.9 (see GCC bug id 57824), coming from
> OptionalTest.cpp, relating to raw string literals in macros not being able
> to encompass multiple lines. The change in question (r351548) was
> introduced 3 days ago, and as far as I can see hasn't flagged up a
build
> bot failure.
> >
> > I'm also seeing a build failure coming from llvm-xray, although in
this
> case it looks like it is actually a change in a header somewhere that is
> causing it. I haven't yet tracked down the revision in question for
this
> (it's possible it's r351701 (and not fixed by r351728) but also
possible
> that something on my machine has got messed up), but removing the contents
> of xray-converter.cpp means I can at least build successfully.
> >
> > Regards,
> >
> > James
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20190122/c945b314/attachment.html>