Somehow when a failure occurred during a parallel build the rule for
.c.o was interrupted mid-step and left an empty .o:
$ ll hist*
16 -rw-r--r-- 1 woods staff - 7671 Feb 23 18:59 hist.d
0 -rw-r--r-- 1 woods staff - 0 Feb 23 18:59 hist.o
456 -rw-r--r-- 1 woods staff - 229544 Feb 23 18:59 hist.o.o
It would be my understanding that this should not be possible.
The rule where this happened is of course this one:
.c.o:
${COMPILE.c} ${.IMPSRC} ${OBJECT_TARGET}
${CTFCONVERT_RUN}
Which basically expands to the following script
cc host.c -o host.o.o
ctfconvert -o host.o host.o.o && rm -f host.o.o
Perhaps though there is a deeper problem with multi-step rules that
generate temporary intermediate files?
In jobs mode make is supposed to feed the whole rule's script to one
invocation of a shell, and in theory that invocation should run to
completion even if some other parallel branch of make encounters an
error.
So, can anyone explain better how this happened?
(Probably this rule should be changed to have the ctfconvert only run if
the compile succeeds, but that shouldn't affect the case I have
discovered here.... Ideally maybe the ctfconvert should be on a
pipeline so that there's no temporary intermediate file necessary, but
that would be more involved given how CTFCONVERT is possibly not always
defined.)
--
Greg A. Woods <gwoods%acm.org@localhost>
Kelowna, BC +1 250 762-7675 RoboHack <woods%robohack.ca@localhost>
Planix, Inc. <woods%planix.com@localhost> Avoncote Farms <woods%avoncote.ca@localhost>
Attachment:
pgp4CGbbZzHM9.pgp
Description: OpenPGP Digital Signature