Repository navigation
BrasilData 0.1.15: replacing a layer no longer removes the one that just landed - #48
elpidioassinaturas wants to merge 1 commit into
Conversation
…ust landed Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
No actionable comments were generated in the recent review. Walkthrough
Merge Risk:
|
Code reviewBugs
Security
Performance
Quality
CLAUDE.md
Plugin packages
Assessment of
Change sizeCompared: the whole PR against main (merge base e61f0c5 to PR head 6244cf8).
|
🧩 Plugin previewOpen in RustGIS: web.rustgis.app's Plugin Manager on this pull request's registry (1 plugin package from this pull request). Install from Browse; RustGIS shows the permissions first.
Unreviewed code: install it only to review this pull request. |
Plugin: BrasilData (
brasildata), elpidioassinaturas/BrasilDataWhat it does: this updates
registry/brasildata.jsonfrom 0.1.14 to 0.1.15, and nothing else.Name (2)while the old name was in use. Before, it removed the old layer first. With a RustGIS that runs a command again when its download lands (feat(plugins): a plugin command whoselayer.addUrlleft a download runs again when it lands, and goes on with the layer (#1202) RustGIS#1341), that run removed the layer that had just landed and ended in an error, and with an area set the old layer could stay beside the new one.How I tested it:
python scripts/build.py --checkwith this entry passes (with Python 3.12): 1 new package, 12 already published. The release's hash matches the entry.Against real services, with the same source as the release:
service.addreads off the frame on the desktop (#1319) RustGIS#1513, in a headless test: the same three cases, with no error;The package from the release itself ran the INPE hotspots case in rustgis-cli 0.5.0, with the same result. Solos could not be run with it: see below.
Not tested: the desktop window; the census tracts; in the web app, the case with an area and the theme that makes its own request.
A known failure, not from this version: Solos can fail with
no layer 'BDIA:pedo_area', on 0.1.14 too and with a plainservice.addwith no plugin. IBGE's WFS lists 10,186 layers, their order changes between answers, and RustGIS reads the first 2,000. When it failed, that layer was at position 10,181. It needs a fix in RustGIS.source.urlis a release asset of the plugin's repository, andsha256is that file's.🤖 Generated with Claude Code
Summary by CodeRabbit