Closed hunleyd closed 2 months ago
from @keithf4
was there ever any intensions of supporting beta releases? If so, I'd say maybe use
current_setting('server_version_num')
instead of trying to rely on parsing that out
@Andersson007 was support for beta/rc intended? any reason to not switch as suggested?
from @keithf4
was there ever any intensions of supporting beta releases? If so, I'd say maybe use
current_setting('server_version_num')
instead of trying to rely on parsing that out@Andersson007 was support for beta/rc intended? any reason to not switch as suggested?
@hunleyd thanks for raising the issue! As long as the suggestion preserves backwards compatibility -i.e. ret values and user playbooks will continue working w/o any change needed- I'm OK with this. Would you like to implement it yourself? If no volunteers, I can try to do it myself, no problem, but volunteers are welcome Please let me know ASAP
@Andersson007 would i like to? yes. will i have time to? probably not until this fri at the earliest
@Andersson007 would i like to? yes. will i have time to? probably not until this fri at the earliest
@hunleyd cool, assigned the issue to you then:) Feel free to re-assign to me if the circumstances change
SUMMARY
When running postgresql_info against one of the beta releases (e.g. PG 17 beta 3), a failure is generated:
This comes from
get_pg_version()
which doesSELECT version()
and then tries to applyr"(\d+)\.(\d+)(?:\.(\d+))?
to said results. Unfortunately, the query in question returns:which means that
fail on lines 997/998
ISSUE TYPE
COMPONENT NAME
postgresql_info
ANSIBLE VERSION
COLLECTION VERSION
CONFIGURATION
OS / ENVIRONMENT
This was against a RHEL 8 based OS
STEPS TO REPRODUCE
Install PG 17beta3, then try to run postgresql_info against it
EXPECTED RESULTS
version properly handled
ACTUAL RESULTS