From: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
---|---|
To: | PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | get_progname and .exe suffix |
Date: | 2004-10-15 14:06:54 |
Message-ID: | 416FD97E.8050906@dunslane.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
My original code for initdb stripped the .exe suffix off the progname
stored, in this snippet of code:
carg0 = xstrdup(argv[0]);
canonicalise_path(carg0);
lastsep = strrchr(carg0, *'/'*);
progname = lastsep ? xstrdup(lastsep + 1) : carg0;
#*if* *defined*(__CYGWIN__) || *defined*(WIN32)
*if* (strlen(progname) > 4 &&
(exe = progname + (strlen(progname) - 4)) &&
stricmp(exe, EXE) == 0)
{
//* strip .exe suffix, regardless of case *//
*exe = *'\0'*;
}
#*endif* //* defined(__CYGWIN__) || defined(WIN32) */
Somewhere along the way, probably when we moved to using a generalised get_progname(), that seems to have been lost. Not only is keeping the .exe suffix ugly, but it breaks the contrib installcheck regression tests on windows (hasn't anyone run these?)
A simple fix for this last problem would be to put some similar code to the above into psql, but something cleaner and applicable to all users of get_progname might be nicer.
Thoughts?
andrew
/
From | Date | Subject | |
---|---|---|---|
Next Message | Dave Page | 2004-10-15 14:12:26 | Re: get_progname and .exe suffix |
Previous Message | Merlin Moncure | 2004-10-15 14:00:48 | Re: libpq and prepared statements progress for 8.0 |