tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Build system race conditions

On Fri 12 Sep 2014 at 09:32:53 +0200, Martin Husemann wrote:
> The issues are generic, random race conditions in the NetBSD build system

Aha! Those may be related then to the issues I've seen whenever I tried
parallel builds on my own machine, that apparently never happened to
anyone else, but which disappeared if I omitted the -j option.

For one of those I found a mail I wrote:

More recently I had a mysterious -j failure where I had to remove a
small subree in my objdir before a non-j make worked again. Looking
through the logfile, I now see this, a bit before a compile failure:

    compile  netpgpverify/bufgap.o
     create  amd/ops_efs.d
     create  amd/ops_mfs.d
    compile  libgroff/uniuni.o
    compile  netpgpverify/digest.o
     create  amd/ops_nfs.d
     create  amd/ops_nfs3.d
    compile  netpgpverify/libverify.o
     create  amd/ops_nullfs.d
     create  amd/ops_pcfs.d
    compile  libgroff/version.o
      build  libgroff/libgroff.a
eval: Cannot fork
eval: Cannot fork
eval: Cannot fork
eval: Cannot fork
eval: Cannot fork
: permission denied
     create  amd/ops_tfs.d
dependall ===> gnu/usr.bin/groff/src/libs/libdriver
     create  libdriver/input.d
     create  libdriver/printer.d
     create  amd/ops_tmpfs.d

and later

       link  grodvi/grodvi
dvi.o: In function `dvi_font::~dvi_font()':
dvi.cpp:(.text+0x18): undefined reference to `font::~font()'
A failure has been detected in another branch of the parallel make

Maybe the parallelism got a bit out of control there with the "cannot
fork" errors?

___ Olaf 'Rhialto' Seibert  -- The Doctor: No, 'eureka' is Greek for
\X/ rhialto/at/    -- 'this bath is too hot.'

Attachment: pgpaYeJWBv2bv.pgp
Description: PGP signature

Home | Main Index | Thread Index | Old Index