Time |
Nick |
Message |
07:55 |
|
BDorsey joined #evergreen |
08:39 |
|
mmorgan joined #evergreen |
09:19 |
|
kworstell-isl joined #evergreen |
09:34 |
Bmagic |
kmlussier++ |
09:34 |
|
dguarrac joined #evergreen |
10:10 |
|
Dyrcona joined #evergreen |
10:29 |
|
sleary joined #evergreen |
10:48 |
|
Christineb joined #evergreen |
10:51 |
Dyrcona |
Results from restoring a db on Pg15: pg_restore: warning: errors ignored on restore: 6 |
10:53 |
Dyrcona |
However, the scroll back only shows 4 errors, two related to the target database and the public schema not existing for the drop statement and the other two are different instances of "pg_restore: error: could not execute query: ERROR: functions in index expression must be marked IMMUTABLE." |
10:53 |
Dyrcona |
Nothing about xml_is_well_formed. |
10:55 |
Dyrcona |
Apparently, we can no long use COALESCE in index creation. |
10:55 |
Dyrcona |
s/long/longer/ |
10:59 |
Dyrcona |
The xml_is_well_formed function exists in my restored database. I wonder if I misunderstood the release note? |
11:00 |
miker |
Dyrcona: tbh, I was a little confused about your comments on xml stuff yesterday ... the current dev docs suggest that xml_is_well_formed is still there, it looked like it was just a matter of how xml2 is installed and what/how it wraps that with xml_valid, extension vs Ye Olde Way. |
11:00 |
miker |
however, and regardless, |
11:00 |
miker |
Dyrcona++ # keeping an eye on those release notes |
11:03 |
Dyrcona |
miker: Yeah. I actually started to think I had misunderstood something yesterday afternoon when reading some of the email discussions shared by jeffdavis. |
11:05 |
Dyrcona |
It looks like we may have to change the definitions of chmm_once_per_paramset and ccmm_once_per_paramset indexes. |
11:05 |
Dyrcona |
As I near as I can tell, Pg doesn't like the use of coalesce in those index definitions. |
11:08 |
|
rfrasur joined #evergreen |
12:08 |
|
jihpringle joined #evergreen |
12:42 |
|
Stompro joined #evergreen |
17:00 |
|
mmorgan left #evergreen |