diff options
| author | Pjotr <wrk@bosa.lan> | 2026-11-08 13:28:15 +0000 |
|---|---|---|
| committer | Pjotr <wrk@bosa.lan> | 2026-11-08 13:28:15 +0000 |
| commit | e5c2aa0ab1e568fe91c63cd5257aee7ff08ebed9 (patch) | |
| tree | 3404fdf0a0c1a0cd34218832905bac0361074b64 /test/python/server_test_case.py | |
| parent | b8fdbb3d60603c07cacf9f406ed1a5c228223829 (diff) | |
| download | genenetwork4-e5c2aa0ab1e568fe91c63cd5257aee7ff08ebed9.tar.gz | |
docs: what the last two commits changed, and the dev/production split
* README.md (Running with cargo): "This deployment runs two instances:
3000 (production) and 3001 (development)" -- the table of ports, binaries
and cache directories, both start-up invocations, and why the caches are
separate while the database, index and auth stub are shared. Also notes
that in-memory state is per instance, so the dev server is not a capacity
test of production.
* docs/www/ai-instruction.md:
- /version now described honestly: git (with -dirty), dirty,
source_digest, database reachable/unreachable -- and that
`unreachable` means stop and report, not retry in a loop.
- error table: 413 (plain text, not the JSON object: over the 64 MiB
body limit), 500 database_error with a connection-loss detail is
transient (retry once), and 409 run_id_in_use pointing at the status
endpoint.
- covariates: case_attribute is the smallest form (nothing posted),
which is also the answer to a 413 from sending a value per individual.
Diffstat (limited to 'test/python/server_test_case.py')
0 files changed, 0 insertions, 0 deletions
