Thread: build error
The fmgroids.h file was incorrectly being generated by shell script
Gen_fmgrtab.sh, so I had to fix the makefile so that instead of the shell script generating the .h
file, perl generated it:
#fmgroids.h fmgrtab.c: Gen_fmgrtab.sh
$(top_srcdir)/src/include/catalog/pg_proc.h # AWK='$(AWK)' $(SHELL) $<
$(top_srcdir)/src/include/catalog/pg_proc.h fmgroids.h fmgrtab.c: Gen_fmgrtab.pl
$(top_srcdir)/src/include/catalog/pg_proc.h perl $< $(top_srcdir)/src/include/catalog/pg_proc.h
in the /usr/local/src/pgsql/src/backend/utils/ directory. The fmgroids.h was being created with the basic
template,but none
of the DEFINES were populated, aside from the wrap-around defines: #ifndef FMGROIDS_H #define FMGROIDS_H
Make check showed no errors and everything seems ok now after that
one fix. This is working off of today's cvs.
> gawk --version
GNU Awk 3.1.5
> uname -a
Linux gnulou 2.6.17-co-0.8.0 #2 PREEMPT Wed Apr 25 22:15:57 CEST 2007
i686 GNU/Linux
> /usr/local/pgsql/bin/psql -U gnulou test
psql (8.5devel)
vanek <vanek@acd.net> writes:
> The fmgroids.h file was incorrectly being generated by shell script
> Gen_fmgrtab.sh,
Uh, it works for everybody else. What platform are you testing on?
(The kernel version isn't a real helpful answer, since it's unlikely
this is a kernel issue.)
While I'm not certain how it would lead to this exact failure, one
thing to check is that your machine's clock is set correctly ---
file timestamps that seem to be in the future can cause "make" to
skip necessary steps.
regards, tom lane
Uh, it works for everybody else. What platform are you testing on?
(The kernel version isn't a real helpful answer, since it's unlikely
this is a kernel issue.)
While I'm not certain how it would lead to this exact failure, one
thing to check is that your machine's clock is set correctly ---
file timestamps that seem to be in the future can cause "make" to
skip necessary steps.
regards, tom lane
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
I'm building on debian etch. The time is set properly. The .h was generated,
just incorrectly, and 'make' didn't skip this step. The problem lies somewhere
downstream from make, possibly with gawk or bash or the bash environment. If
you don't think it's a problem, it's not a problem. I just thought I'd report
it because I know a few others have encountered this error, too.
vanek wrote: > Uh, it works for everybody else. What platform are you testing on? > (The kernel version isn't a real helpful answer, since it's unlikely > this is a kernel issue.) > > While I'm not certain how it would lead to this exact failure, one > thing to check is that your machine's clock is set correctly --- > file timestamps that seem to be in the future can cause "make" to > skip necessary steps. > > regards, tom lane > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > I'm building on debian etch. The time is set properly. The .h was > generated, > just incorrectly, and 'make' didn't skip this step. The problem lies > somewhere > downstream from make, possibly with gawk or bash or the bash > environment. If > you don't think it's a problem, it's not a problem. I just thought I'd > report > it because I know a few others have encountered this error, too. > There are quite a few debian members of the buildfarm that are not having any problem - see <http://www.pgbuildfarm.org/cgi-bin/show_status.pl>. So the question is "why are you having a problem they aren't having?" You haven't given us nearly enough information to be able to guess why it should be so. You need to show us a) the output from configure and b) the complete output from make (both stdout and stderr). c) the configure options you are using. Are you having this problem with a fresh checkout? cheers andrew