Changes between Initial Version and Version 1 of Ticket #5545, comment 11


Ignore:
Timestamp:
10/06/23 08:13:53 (15 months ago)
Author:
strk

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #5545, comment 11

    initial v1  
    66This is usually easy when `SELECT postgis_extensions_upgrade()` is able to package those raster functions, as then you just drop the extension, but you stumbled upon a bug which prevents packaging of those functions (now filed as #5550), and is a bug that's most likely only encountered when using a sandboxed system, because we do have regression testing in place and all bots are happy with the repackaging of raster (see #5212 for a ticket to request improving the CI coverage to add this case).
    77
    8 The above said, I'm not sure there's much more to do in this ticket. The main problem was that the upgrade script from 3.2.3 to 3.3.2 failed to drop the st_orderingequals, but we didn't yet confirm it was due to sandboxing. I'm not sure your invocation of {{{
     8The above said, I'm not sure there's much more to do in this ticket. The main problem was that the upgrade script from 3.2.3 to 3.3.2 failed to drop the st_orderingequals, but we didn't yet confirm it was due to sandboxing. I'm not sure your invocation of
     9{{{
    910SELECT _postgis_drop_function_if_needed
    1011        (
    … …  
    1415}}}
    1516
     17
    1618Did something or not, maybe you can re-run after {{{ set client_min_messages to DEBUG; }}} to know more ? What we are interested in is WHY that function wasn't removed when you initially upgraded.
    1719