Hi Eric, * Eric Blake wrote on Mon, Oct 06, 2008 at 04:43:45PM CEST: > Eric Blake <ebb9 <at> byu.net> writes: > > > > > +if test $at_jobs -ne 1 && > > > + rm -f "$at_job_fifo" && > > > + ( mkfifo "$at_job_fifo" ) 2>/dev/null && > > > + exec AT_JOB_FIFO_FD<> "$at_job_fifo" > > > > Is exec n<> portable? Yes, POSIX specifies it, but I haven't seen it used > > anywhere else in autoconf. I guess we commit it now, and if it is not > > portable, we wrap it inside an eval alongside the mkfifo check as a reason > > to skip the parallel testing. > > Unfortunately, it is NOT portable.
Thanks for checking; and: sigh. That will require quite a bit of retesting, I'm afraid. Oh well. Can we work around it for now by detecting the Cygwin failure easily (thus disabling parallel tests there)? > As a result, the parallel tests invariably fail: > +./micro-suite: line 1703: read: read error: 0: Communication error on send Interesting, it didn't do that when I tested it (quite a while ago). > Can we work on refactoring this into two fd's both visiting the same FIFO, > one > for writing tokens, and the other for reading tokens, rather than using <>? Yes, that's one idea I'd like to work on (or be beaten to ;-) Unless you plan to cut a release soon, it wouldn't be my very next thing to do though. Thanks, Ralf
