Opened 2 years ago
Closed 2 years ago
#34760 closed Cleanup/optimization (fixed)
Drop support for SQLite < 3.27.0
Reported by: | Mariusz Felisiak | Owned by: | Mariusz Felisiak |
---|---|---|---|
Component: | Database layer (models, ORM) | Version: | 4.2 |
Severity: | Normal | Keywords: | |
Cc: | Claude Paroz | Triage Stage: | Accepted |
Has patch: | yes | Needs documentation: | no |
Needs tests: | no | Patch needs improvement: | no |
Easy pickings: | no | UI/UX: | no |
Description
SQLite 3.24+ supports updating rows on constraint or uniqueness errors during INSERT
.
SQLite 3.25+ supports window functions and renaming columns.
SQLite 3.26+ support DDL operations requiring references rename in a transaction.
Ubuntu Focal ships with SQLite 3.31.1 (which will still by supported by Django).
Debian Buster ships with 3.27.2 and will EOL in June 2024.
Python 3.8 ships with 3.28.0.
SQLite 3.26.0 was released in December 2018. SQLite version support seems like a similar situation as GEOS libraries which we generally support about 5 years after released.
Change History (7)
comment:1 by , 2 years ago
Owner: | changed from | to
---|---|
Status: | new → assigned |
follow-up: 3 comment:2 by , 2 years ago
Triage Stage: | Unreviewed → Accepted |
---|
follow-up: 4 comment:3 by , 2 years ago
Replying to Natalia Bidart:
Sounds good, though following the rationale I'm not sure I understand why we wouldn't drop support for < 3.27?
3.27 will not be 5 years old, when Django 5.0 is released.
comment:4 by , 2 years ago
Replying to Mariusz Felisiak:
Replying to Natalia Bidart:
Sounds good, though following the rationale I'm not sure I understand why we wouldn't drop support for < 3.27?
3.27 will not be 5 years old, when Django 5.0 is released.
Right, but 3.26 will be, so I was wondering if we should also drop 3.26. From the ticket title I understood you mean drop every version less than 3.26? Or did you mean <= 3.26
"?
comment:5 by , 2 years ago
Summary: | Drop support for SQLite < 3.26.0 → Drop support for SQLite < 3.27.0 |
---|
Agreed, let's bump to 3.27.0.
Sounds good, though following the rationale I'm not sure I understand why we wouldn't drop support for < 3.27?