Re: Call for Foundation-supported Project Ideas

From: Steve Kargl <sgk_at_troutmask.apl.washington.edu>
Date: Mon, 29 Nov 2021 04:36:12 UTC
On Sun, Nov 28, 2021 at 10:17:00PM -0500, Ed Maste wrote:
> On Sun, 28 Nov 2021 at 19:37, Steve Kargl
> <sgk@troutmask.apl.washington.edu> wrote:
> >
> > It's certainly not the latest and greatest,
> > CPU: Intel(R) Core(TM)2 Duo CPU     T7250  @ 2.00GHz (1995.04-MHz K8-class CPU)
> 
> If you're content to use a compiler from a package you can save a lot
> of time by building with `CROSS_TOOLCHAIN=llvm13` and
> `WITHOUT_TOOLCHAIN=yes`. Or, instead of WITHOUT_TOOLCHAIN perhaps
> `WITHOUT_CLANG=yes`, `WITHOUT_LLD=yes` and `WITHOUT_LLDB=yes`.
> 
> Cirrus-CI builds CROSS_TOOLCHAIN and WITHOUT_TOOLCHAIN to reduce build
> time significantly, given that the vast majority of CI build/test runs
> are not targeting compiler changes.

And therein lies the rub.  Cirrus-CI skirts the issue with the
slowing down of builds, so developers don't see the problem.  I
suspect developers also tend to use ccache or meta_mode.  Reported
"buildworld times" are not really "buildworld times" but more like
partial "buildworld times".

Unfortunately, I've wasted a weekend trying to get an up-to-date
to chase down what may be a bug in (1) how the kernel delivers
signals to a thread; (2) a problem with libthr; or a problem with
asynchronous IO in gfortran.  My REAL buildworld will finish
eventually, so I'l be able to look more closely at the issue.

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103444

-- 
Steve