| commit | f02154652d1a83d80f4bee93b58ecdbd8c33378b | |
| author | Tom Lane <tgl@sss.pgh.pa.us> | |
| Sat, 30 Sep 2023 00:20:57 +0000 (29 20:20 -0400) | ||
| committer | Tom Lane <tgl@sss.pgh.pa.us> | |
| Sat, 30 Sep 2023 00:21:10 +0000 (29 20:21 -0400) | ||
| tree | d3ec8e31066fe744607ba7f97b3f29a0ce545fff | treesnapshot (tar.gz zip) |
| parent | 0cfc231aa19d80b31e37422a968d51fa8bcb53a9 | commitdiff |
Remove environment sensitivity in pl/tcl regression test.
Add "-gmt 1" to our test invocations of the Tcl "clock" command,
so that they do not consult the timezone environment. While it
doesn't really matter which timezone is used here, it does
matter that the command not fall over entirely. We've now
discovered that at least on FreeBSD, "clock scan" will fail if
/etc/localtime is missing. It seems worth making the test
insensitive to that.
Per Tomas Vondras' buildfarm animal dikkop. Thanks to
Thomas Munro for the diagnosis.
Discussion: https://postgr.es/m/316d304a-1dcd-cea1-3d6c-27f794727a06@enterprisedb.com
Add "-gmt 1" to our test invocations of the Tcl "clock" command,
so that they do not consult the timezone environment. While it
doesn't really matter which timezone is used here, it does
matter that the command not fall over entirely. We've now
discovered that at least on FreeBSD, "clock scan" will fail if
/etc/localtime is missing. It seems worth making the test
insensitive to that.
Per Tomas Vondras' buildfarm animal dikkop. Thanks to
Thomas Munro for the diagnosis.
Discussion: https://postgr.es/m/316d304a-1dcd-cea1-3d6c-27f794727a06@enterprisedb.com
| src/pl/tcl/expected/pltcl_setup.out | diffblobblamehistory | |
| src/pl/tcl/sql/pltcl_setup.sql | diffblobblamehistory |