| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> |
| Cc: | Joseph Brenner <doomvox(at)gmail(dot)com>, pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Select works only when connected from login postgres |
| Date: | 2016-12-04 17:31:40 |
| Message-ID: | 25727.1480872700@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general pgsql-hackers |
Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> writes:
> My suspicion is that when you did the source build you got some cross
> contamination of libraries.
That's the best theory I can come up with either, although libpq's
APIs haven't really changed in any non-backwards-compatible fashion
in years. I could imagine a newer psql flat-out crashing because it
tries to call some libpq function that doesn't exist in an older
libpq, but that's not what we're seeing here. It's really weird.
A couple of tests that might help narrow things down:
1. In the server configuration, enable log_connections,
log_disconnections, and log_statement = 'all'. Does anything show up in
the log when you connect with the broken psql and issue commands?
2. If you issue commands that span multiple lines --- unclosed left
parens, multiline string literals, missing semicolons --- does the
psql prompt change to match?
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2016-12-04 17:39:17 | Re: count(*) in binary mode returns 0 |
| Previous Message | Joseph Brenner | 2016-12-04 17:14:58 | Re: Select works only when connected from login postgres |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2016-12-04 17:51:49 | Re: Select works only when connected from login postgres |
| Previous Message | Joseph Brenner | 2016-12-04 17:14:58 | Re: Select works only when connected from login postgres |