Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Tom Lane
Decibel! <[EMAIL PROTECTED]> writes: > Perhaps if we don't have log_timezone set yet we could output something > such as 'LOCALTIME' so that it was obvious. Or maybe output whatever $TZ > is set to? I think you're confused. This is not a matter of knowing what to print for the timezone part, it's

Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Decibel!
On Sat, Aug 04, 2007 at 12:46:32PM -0400, Tom Lane wrote: > Decibel! <[EMAIL PROTECTED]> writes: > > Something else to consider... normally you'd have to have a pretty > > serious condition to run into this in normal usage, right? (I doubt > > there's many folks that use any debug level, let alone

Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Tom Lane
Decibel! <[EMAIL PROTECTED]> writes: > Something else to consider... normally you'd have to have a pretty > serious condition to run into this in normal usage, right? (I doubt > there's many folks that use any debug level, let alone 3) I think that > gives us more flexibility. Yeah. This whole is

Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Decibel!
On Sat, Aug 04, 2007 at 12:12:46PM -0400, Tom Lane wrote: > There's a value judgment involved here: is it better to print timestamps > in GMT (which the user won't be expecting) than to not print them at all? > I can see arguments on both sides of that. As long as it's clear that the times are in

Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Tom Lane
Magnus Hagander <[EMAIL PROTECTED]> writes: > Any way to tweak this place and have it log in GMT instead, when > log_timezone is NULL? I think this is just a variant of my #2. The main risk factor IMHO is that the first call is going to invoke tzload() for the GMT zone (pgtz does not seem to spec

Re: [HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Magnus Hagander
Tom Lane wrote: > I came across a problem with the log_timezone patch. Given these > conditions: > * log_min_messages is at least debug3 > * log_line_prefix includes %t or %m > * log_timezone is not explicitly set in postgresql.conf > the postmaster will dump core during startup, because it tries

[HACKERS] Problem with log_timezone not being set during early startup

2007-08-04 Thread Tom Lane
I came across a problem with the log_timezone patch. Given these conditions: * log_min_messages is at least debug3 * log_line_prefix includes %t or %m * log_timezone is not explicitly set in postgresql.conf the postmaster will dump core during startup, because it tries to format an elog message us