Skip to content

Commit bc7885b

Browse files
committed
Work around portability issue with newer versions of mktime().
Recent glibc versions have made mktime() fail if tm_isdst is inconsistent with the prevailing timezone; in particular it fails for tm_isdst = 1 when the zone is UTC. (This seems wildly inconsistent with the POSIX-mandated treatment of "incorrect" values for the other fields of struct tm, so if you ask me it's a bug, but I bet they'll say it's intentional.) This has been observed to cause cosmetic problems when pg_restore'ing an archive created in a different timezone. To fix, do mktime() using the field values from the archive, and if that fails try again with tm_isdst = -1. This will give a result that's off by the UTC-offset difference from the original zone, but that was true before, too. It's not terribly critical since we don't do anything with the result except possibly print it. (Someday we should flush this entire bit of logic and record a standard-format timestamp in the archive instead. That's not okay for a back-patched bug fix, though.) Also, guard our only other use of mktime() by having initdb's build_time_t() set tm_isdst = -1 not 0. This case could only have an issue in zones that are DST year-round; but I think some do exist, or could in future. Per report from Wells Oliver. Back-patch to all supported versions, since any of them might need to run with a newer glibc. Discussion: https://postgr.es/m/CAOC+FBWDhDHO7G-i1_n_hjRzCnUeFO+H-Czi1y10mFhRWpBrew@mail.gmail.com
1 parent 47d5781 commit bc7885b

File tree

2 files changed

+27
-5
lines changed

2 files changed

+27
-5
lines changed

src/bin/initdb/findtimezone.c

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -195,6 +195,7 @@ build_time_t(int year, int month, int day)
195195
tm.tm_mday = day;
196196
tm.tm_mon = month - 1;
197197
tm.tm_year = year - 1900;
198+
tm.tm_isdst = -1;
198199

199200
return mktime(&tm);
200201
}

src/bin/pg_dump/pg_backup_archiver.c

Lines changed: 26 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -3712,7 +3712,6 @@ ReadHead(ArchiveHandle *AH)
37123712
vmin,
37133713
vrev;
37143714
int fmt;
3715-
struct tm crtm;
37163715

37173716
/*
37183717
* If we haven't already read the header, do so.
@@ -3780,6 +3779,8 @@ ReadHead(ArchiveHandle *AH)
37803779

37813780
if (AH->version >= K_VERS_1_4)
37823781
{
3782+
struct tm crtm;
3783+
37833784
crtm.tm_sec = ReadInt(AH);
37843785
crtm.tm_min = ReadInt(AH);
37853786
crtm.tm_hour = ReadInt(AH);
@@ -3788,12 +3789,32 @@ ReadHead(ArchiveHandle *AH)
37883789
crtm.tm_year = ReadInt(AH);
37893790
crtm.tm_isdst = ReadInt(AH);
37903791

3791-
AH->archdbname = ReadStr(AH);
3792-
3792+
/*
3793+
* Newer versions of glibc have mktime() report failure if tm_isdst is
3794+
* inconsistent with the prevailing timezone, e.g. tm_isdst = 1 when
3795+
* TZ=UTC. This is problematic when restoring an archive under a
3796+
* different timezone setting. If we get a failure, try again with
3797+
* tm_isdst set to -1 ("don't know").
3798+
*
3799+
* XXX with or without this hack, we reconstruct createDate
3800+
* incorrectly when the prevailing timezone is different from
3801+
* pg_dump's. Next time we bump the archive version, we should flush
3802+
* this representation and store a plain seconds-since-the-Epoch
3803+
* timestamp instead.
3804+
*/
37933805
AH->createDate = mktime(&crtm);
3794-
37953806
if (AH->createDate == (time_t) -1)
3796-
pg_log_warning("invalid creation date in header");
3807+
{
3808+
crtm.tm_isdst = -1;
3809+
AH->createDate = mktime(&crtm);
3810+
if (AH->createDate == (time_t) -1)
3811+
pg_log_warning("invalid creation date in header");
3812+
}
3813+
}
3814+
3815+
if (AH->version >= K_VERS_1_4)
3816+
{
3817+
AH->archdbname = ReadStr(AH);
37973818
}
37983819

37993820
if (AH->version >= K_VERS_1_10)

0 commit comments

Comments
 (0)