- 27 Oct, 2021 1 commit
-
-
Kevin Morris authored
Whitelisted keys: AURSID, AURTZ, AURLANG Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 26 Oct, 2021 1 commit
-
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 25 Oct, 2021 2 commits
-
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Two new options have been added: - [devel] commit_url - URL including an %s format specifier that can be used to link to a webpage for the commit. - [devel] commit_hash - HEAD's commit hash (produced via `git rev-parse HEAD`) If a `[devel] commit_hash` is configured, a link to the commit based on `[devel] commit_url` will be displayed in the aurweb footer in the form: `HEAD@<commit_hash>`. If no `[devel] commit_url` is configured, a non-linked hash will be displayed. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 24 Oct, 2021 7 commits
-
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
The default recursion limit used by Docker's archlinux:base-devel Python package becomes problematic in some cases when running tests against our FastAPI application using starlette.testclient.TestClient (aliased to fastapi.testclient.TestClient). starlette ends up with test failures because it exceeds the recursion limit, but this only happens when using the `TestClient`. When the ASGI servers are run, this is not an issue and so in that case, the recursion limit has not been touched. This change uses a `TEST_RECURSION_LIMIT` environment variable to modify the recursion limit of the FastAPI application. This variable is, by default, only supplied when running pytests in Docker, but can be force-supplied by the user. TEST_RECURSION_LIMIT=10000 has been added to `.env` and `.gitlab-ci.yml`. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Changes: - `via` is not required in FastAPI. We deduce the involved requests via their PackageBaseName / MergeBaseName columns and set them to Accepted when merged. - When erroneous input is given, the error is now presented on the merge page instead of sending the user to the pkgbase page. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Improvements: - Package deletion now creates a PackageRequest on behalf of the deleter if one does not yet exist. - All package deletions are now logged to keep track of who did what. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 23 Oct, 2021 6 commits
-
-
Kevin Morris authored
This merge requires production users to specify an host directory to bind as the git repository within Docker containers. This means that a repository can be shared between host and container, so that the repository does not need to be managed within Docker alone. New environment variables: - GIT_DATA_DIR: Path to aur.git repository that is bind mounted Do note, this variable only takes affect when users run production Docker services, by supplying: $ docker-compose -f docker-compose.yml -f docker-compose.prod.yml ...
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
This deprecation warning is not fixed in a tagged release yet. Ignoring it for now; it has nothing to do with user code. Upstream bug at https://bugs.python.org/issue45097 Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
There were some test failures caused by problematic dependency versioning, most likely to to the seriously braindead pyproject.toml config for deps that previously existed. This commit defines python version >=3.9<3.10 for our working Python version and provides updated deps (to latest). I believe the bug was originally caused by the fact that we had no python dependency defined, allowing poetry to resolve dependencies incorrectly for what we intended. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 22 Oct, 2021 1 commit
-
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
- 21 Oct, 2021 18 commits
-
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Steven Guikal authored
Signed-off-by:
Steven Guikal <void@fluix.one>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
This change exposes the uwsgi daemon we use for cgit on: - PHP: docker-host:13000 - FastAPI: docker-host:13001 These ports can then be used to take advantage of cgit on a production server that hosts nginx in front of Docker. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Additionally, clone-prefix is now configurable via environment variables: - CGIT_CLONE_PREFIX_PHP - CGIT_CLONE_PREFIX_FASTAPI These vars can be used by production to customize the clone prefix. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
New configuration options: - `[ratelimit] cache` - A boolean indicating whether we should use configured cache (1) or database (0) for ratelimiting. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Additionally, added RPC.error, which produces an RPC-compatible error based on the version passed during construction. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
With this change, we provide a wrapper to `logging.getLogger` in the `aurweb.logging` module. Modules wishing to log using logging.conf should get their module-local loggers by calling `aurweb.logging.getLogger(__name__)`, similar to `logging.getLogger`, this way initialization with logging.conf is guaranteed. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
When using this input on `live` as a TU, the field is not taken into account. Tried with no action and with the Delete Packages action, which ended up deleting the packages but not merging into the given target. So, this commit removes that input from the page. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
-
Kevin Morris authored
This reworks the base implementation of the RPC to use a class called RPC for handling of requests. Took a bit of a different approach than PHP in terms of exposed methods, but it does end up achieving the same goal, with one additional error: "Request type '{type}' is not yet implemented." For FastAPI development, we'll stick with: - If the supplied 'type' argument has an alias mapping in RPC.ALIASES, we convert the type argument over to its alias before doing anything. Example: 'info' is aliased to 'multiinfo', so when a user requests type=info, it is converted to type=multiinfo. - If the type does not exist in RPC.EXPOSED_TYPES, the following error is produced: "No request type/data specified." - If the type **does** exist in RPC.EXPOSED_TYPES, but does not have an implemented `RPC._handle_{type}_type` function, the following error is produced: "Request type '{type}' is not yet implemented." Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
This feature of RPC is required to take advantage of javascript typeahead. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
-
Kevin Morris authored
-
- 19 Oct, 2021 4 commits
-
-
Steven Guikal authored
Closes #134 Signed-off-by:
Steven Guikal <void@fluix.one>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Signed-off-by:
Kevin Morris <kevr@0cost.org>
-
Kevin Morris authored
Previously, we passed the straight up request type instance from SQLAlchemy and had a .title() function that was transparently treating the instance the same as the instance's Name in terms of notify.py's use of it. This commit removes that transparent behavior; it was not actually intended. Signed-off-by:
Kevin Morris <kevr@0cost.org>
-