Re: pg_regress in C

"Magnus Hagander" <mha(at)sollentuna(dot)net> writes:
> Per discussion at the conference:
> In order to run the regression tests on Windows without msys, pg_regress
> needs to be reimplemnted in C.

Patch committed after significant further work. As committed,
pg_regress.c is pretty nearly an exact replacement for the shell script;
the only significant deviation is that the --temp_install switch's
argument is required not optional. (This is because our homegrown
version of getopt_long doesn't allow optional arguments. Maybe that
should be fixed sometime.)

There is one possibly important loose end: the shell script makes an
effort to catch signals and shut down the temp postmaster before
exiting, while there's no such code in the C version. I'm not sure
if it's necessary. At least on my machine, if you type control-C while
the tests are running then the kernel sends SIGINT to everything that's
part of the terminal's process group, which will include the postmaster
--- so the shutdown happens anyway. I have no idea if that'll work on
Windows... One reason I didn't try to do this is I'm a bit hesitant to
write a signal handler that does anything as interesting as a system()
call, which would seem to be necessary to duplicate what the shell
script did. Comments?

regards, tom lane