| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 18 hours | update changelogHEADmaster | Daniel Schadt | |
| 18 hours | Merge branch 'fs-transactions' | Daniel Schadt | |
| 35 hours | move connection to context manager (fix deadlock?) | Daniel Schadt | |
| Postgres doesn't like us altering a table while a transaction is open, and it seems that by not closing those connections in the update scripts, we provoke a deadlock in the CI. Not sure what changed now (because it worked previously), but I hope this fixes it by properly closing connections as soon as they are no longer needed. | |||
| 36 hours | remove gitlab CI | Daniel Schadt | |
| We're running out of compute minutes every month anyway, and I have set up my own woodpecker server. | |||
| 37 hours | fix type of create_data_folders | Daniel Schadt | |
| 37 hours | ensure tracks/users folders exist | Daniel Schadt | |
| Since we don't use parents=True anymore to create the folders (which I think is good), we now need to ensure that those exist. So 1. when fietsboek starts up, we create those folders, and 2. when we delete them in the tests, we recreate them. | |||
| 38 hours | fix formatting | Daniel Schadt | |
| 38 hours | remove TrackDataDir context manager use in edit | Daniel Schadt | |
| 38 hours | fix types in fstrans | Daniel Schadt | |
| 38 hours | actually add txn parameter to UserDataDir | Daniel Schadt | |
| 38 hours | better logging of exceptions during rollback | Daniel Schadt | |
| 39 hours | more doc strings | Daniel Schadt | |
| 42 hours | initial filesystem transactions | Daniel Schadt | |
| This should solve issues that arise when exceptions occur during upload. Hooks into the transaction/pyramid_tm machinery. | |||
| 4 days | move pyramid to Kingdread/pyramid fork | Daniel Schadt | |
| No more pkg_resources warnings | |||
| 4 days | fail gracefully for __VERSION__ | Daniel Schadt | |
| 2025-12-07 | Merge branch 'pdf' | Daniel Schadt | |
| 2025-11-29 | speed up track adding in test_browse | Daniel Schadt | |
| 2025-11-29 | add smoke test for pdf generation | Daniel Schadt | |
| This doesn't check any content of the PDF, but ensures that running typst does not produce an error. | |||
| 2025-11-29 | ignore matploblib debug logs in tests | Daniel Schadt | |
| matplotlib procudes a lot of lines for the font manager, which really clutter the log output for everything else. | |||
| 2025-11-29 | fix TrackWithMetadata attribute access | Daniel Schadt | |
| 2025-11-29 | adjust image size in pdf | Daniel Schadt | |
| 2025-11-25 | add unit tests for typst escaping | Daniel Schadt | |
| 2025-11-25 | add button to download PDF | Daniel Schadt | |
| 2025-11-22 | remove unneeded import | Daniel Schadt | |
| 2025-11-22 | add test for xml_escape | Daniel Schadt | |
| 2025-11-22 | remove uneeded function | Daniel Schadt | |
| 2025-11-22 | remove unneeded import | Daniel Schadt | |
| 2025-11-22 | don't offer brotli content encoding | Daniel Schadt | |
| That was good while we still had files saved brotli-compressed, but now it adds more overhead. | |||
| 2025-11-22 | speed up track xml serialization | Daniel Schadt | |
| The comment explains it. | |||
| 2025-11-20 | fix types | Daniel Schadt | |
| 2025-11-20 | add height profile to PDF | Daniel Schadt | |
| 2025-11-20 | add docstrings & fix lint | Daniel Schadt | |
| 2025-11-20 | add translation strings for PDF | Daniel Schadt | |
| 2025-11-20 | initial PDF generation | Daniel Schadt | |
| This adds initial functionality to render PDF overviews of tracks. I was pondering to use reportlab and do it completely in Python, but the effort of doing proper layouting seemed to much. Add to that the fact that Typst has much nicer typesetting, it seems like a no-brainer to use it. | |||
| 2025-11-16 | fix lint | Daniel Schadt | |
| 2025-11-16 | enable foreign key checks in SQLite | Daniel Schadt | |
| 2025-11-16 | properly clean up track cache in test | Daniel Schadt | |
| 2025-11-16 | further speed up tests | Daniel Schadt | |
| A lot of time in test_browse was spent retrieving path points and computing metadata. Why not save it while we still have the path ready? | |||
| 2025-11-16 | fix lint | Daniel Schadt | |
| 2025-11-16 | speed up track insertion | Daniel Schadt | |
| It turns out that adding multiple thousands of track points via the ORM is quite slow, which is especially noticable in the browse_paged test (as 50 tracks are added there). This provides the fast_set_path method, which goes directly to the database, utilizing the executemany() capability. On the CI, the gpx-influx decreased test runtime from ~4min to ~18min. This change should make it much faster again. | |||
| 2025-11-15 | refine woodpecker config | Daniel Schadt | |
| 2025-11-14 | enable woodpecker CI | Daniel Schadt | |
| 2025-11-14 | make lon/lat non-nullable | Daniel Schadt | |
| I want to have more columns non-nullable, but it's a pain to change in SQLite. We can still try to be careful with new columns, though, and use nullable=False more often when appropriate. While this changes the schema without a new migration, I think for this case it is fine. | |||
| 2025-11-14 | fix SQLAlchemy types | Daniel Schadt | |
| See https://docs.sqlalchemy.org/en/20/changelog/whatsnew_20.html#whatsnew-20-orm-declarative-typing The Mapped[] annotations help mypy to find the right types for the instance attributes. Some of the mapped_column() definitions are superfluous, but I think it's nice to have them explicit. | |||
| 2025-11-14 | be graceful if users/tracks dirs not found | Daniel Schadt | |
| This may happen on a fresh instance, when you navigate to the admin view. | |||
| 2025-11-14 | Merge branch 'gpx-influx' | Daniel Schadt | |
| 2025-11-14 | add missing return statement | Daniel Schadt | |
| 2025-11-13 | check DB connectivity before updating | Daniel Schadt | |
| 2025-11-12 | respect track date when downgrading | Daniel Schadt | |
| 2025-11-11 | fix lint | Daniel Schadt | |
 |
index : fietsboek | |
| A web-based GPX track tracker | Daniel |
| aboutsummaryrefslogtreecommitdiff |