Overview
Request 1102089 accepted
I updated main_version to 3.3 and sfcgal-devel >= 1.4.1
We already had
BuildRequires: libgeos-devel >= 3.7.0
BuildRequires: libproj-devel >= 6.0.0
That should be sufficient to link new enough versions and require them at runtime via rpm's magic.
For the changelog, I already had to drop > 50% of 3.3.3 entries, because they were identical to 3.2.5's, so I did not want to go further down the history. Feel free to adapt it to your liking.
- update to 3.3.4
* Bug Fixes and Enhancements *
- #5450, Fix macro expansion recursion on powerpc architectures.
(Bas Couwenberg)
- #5395, [postgis_topology] Allow unprivileged user dump
of database with topology enabled
- #5394, [postgis_topology] Improve robustness of finding
distinct vertices in topology edges (Sandro Santilli)
- #5403, [postgis_topology] Fix ValidateTopology(bbox) without
topology in search_path (Sandro Santilli)
- #5410, [postgis_raster] ST_Value bilinear resample,
don't throw an error if Band has no NODATA value
(Regina Obe)
- #5385, Postgres malloc assertion fail when
using pg_cancel_backend with ST_AsMVT
(Regina Obe, Paul Ramsey)
- #5452, Updated README.postgis for apt-based systems
(Jelte Fennema)
* Breaking Changes *
- #5385, Take out interruptability of ST_AsMVT
as it causes backend crash under intense conditions
(Regina Obe, Paul Ramsey)
PostGIS 3.3.3
* Bug Fixes and Enhancements *
- #5303, Don't try removing closed edges endpoint from
RemoveUnusedPrimitives (Sandro Santilli)
- #5289, Fix misleading message in RemoveUnusedPrimitives about
doubly connected edges healing (Sandro Santilli)
- #5298, Fix CopyTopology exception with hierarchical layers (Sandro Santilli)
- #5299, Corrections to address_standardizer_data_us lex (Regina Obe)
- Created by bmwiedemann
- In state accepted
-
Package maintainers:
bruno_friedmann,
dstoecker, and
odassau
- Supersedes 1101964
I'd think main_version needs to be increased to 3.3. Also your changelog is indicating, that 3.3.3 comes after 3.2.5, but there is lots of other stuff inbetween. You should at least mention that there were such changes.
Also geos needs to be >= 3.6 now, proj >= 5.2 and sfcgal >= 1.4.1
Request History
bmwiedemann created request
I updated main_version to 3.3 and sfcgal-devel >= 1.4.1
We already had
BuildRequires: libgeos-devel >= 3.7.0
BuildRequires: libproj-devel >= 6.0.0
That should be sufficient to link new enough versions and require them at runtime via rpm's magic.
For the changelog, I already had to drop > 50% of 3.3.3 entries, because they were identical to 3.2.5's, so I did not want to go further down the history. Feel free to adapt it to your liking.
- update to 3.3.4
* Bug Fixes and Enhancements *
- #5450, Fix macro expansion recursion on powerpc architectures.
(Bas Couwenberg)
- #5395, [postgis_topology] Allow unprivileged user dump
of database with topology enabled
- #5394, [postgis_topology] Improve robustness of finding
distinct vertices in topology edges (Sandro Santilli)
- #5403, [postgis_topology] Fix ValidateTopology(bbox) without
topology in search_path (Sandro Santilli)
- #5410, [postgis_raster] ST_Value bilinear resample,
don't throw an error if Band has no NODATA value
(Regina Obe)
- #5385, Postgres malloc assertion fail when
using pg_cancel_backend with ST_AsMVT
(Regina Obe, Paul Ramsey)
- #5452, Updated README.postgis for apt-based systems
(Jelte Fennema)
* Breaking Changes *
- #5385, Take out interruptability of ST_AsMVT
as it causes backend crash under intense conditions
(Regina Obe, Paul Ramsey)
PostGIS 3.3.3
* Bug Fixes and Enhancements *
- #5303, Don't try removing closed edges endpoint from
RemoveUnusedPrimitives (Sandro Santilli)
- #5289, Fix misleading message in RemoveUnusedPrimitives about
doubly connected edges healing (Sandro Santilli)
- #5298, Fix CopyTopology exception with hierarchical layers (Sandro Santilli)
- #5299, Corrections to address_standardizer_data_us lex (Regina Obe)
dstoecker accepted request
Simply add a comment like "Higher version upgrade, see ...URL... for the full log." or something similar. It's not necessary to fiddle with the details, only make clear what's happening.
Hmmpf, not sure about the correct wording. "Major version" is wrong as that means the first number changes. "Minor version" would be correct, but that makes no real sense. So I did use "Higher version" which is a bit strange, but better than Minor. Maybe you find a better text.
@bruno_friedmann, @dstoecker, @odassau: review reminder