Re: llvm10 build failure on Rpi3

http://www.zefox.org/~bob/poudriere/bulk.log shows that you had
the RPi3b (whatever variant) with the following builders (jobs)
going in parallel as of whe llvm10 reported the failure:

. . .
[00:03:20] Building 81 packages using 4 builders
. . .
[00:07:07] [04] [00:00:00] Building lang/rust | rust-1.53.0
. . .
[04:10:52] [01] [00:00:00] Building devel/llvm10 | llvm10-10.0.1_5

After that various things build in builders 02 and 03 as well
but:

. . .
[06:53:56] [02] [00:41:42] Finished print/harfbuzz-icu | harfbuzz-icu-2.8.1: Success
[09:01:25] [03] [02:18:33] Finished multimedia/ffmpeg | ffmpeg-4.4_2,1: Success
[18:53:17] [01] [14:42:25] Finished devel/llvm10 | llvm10-10.0.1_5: Failed: build
[18:54:05] [01] [14:43:13] Skipping www/chromium | chromium-91.0.4472.114_1: Dependent port devel/llvm10 | llvm10-10.0.1_5 failed
[18:54:05] [01] [14:43:13] Skipping graphics/mesa-libs | mesa-libs-20.2.3_1: Dependent port devel/llvm10 | llvm10-10.0.1_5 failed

(The file context does not allow figuring out how many
processes per builder/jobs were allowed or active, beyond
it being at least 1.)

Trying to build lang/rust and devel/llvm10 in parallel on
a low RAM environment is not likely to ever work well, much
less having other things also building at times especially
if multiple processes are allowed to be active per builder
(job). lang/rust is more resource intensive than devel/llvm10
as I understand.

Running them one at a time in sequence is more likely
appropriate in such a context.

===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)