conda_mirror.py: set additional parameters from config-file... #74
Conversation
…ping ability to override from command line
Codecov Report
@@ Coverage Diff @@
## master #74 +/- ##
==========================================
- Coverage 93.89% 92.11% -1.78%
==========================================
Files 2 2
Lines 262 279 +17
==========================================
+ Hits 246 257 +11
- Misses 16 22 +6
Continue to review full report at Codecov.
|
|
Seems like the PR was a bit premature - the arguments that have default values set to something that does not evaluate to False are not overridden by values in the config file. Like temp_directory. |
conda_mirror.py: added command line swith: --no-validate-target
Handle download errors
|
After some interesting merges and some fixup, the change/pr should be good to go now. |
| # ignore values that can only be given on command line | ||
| (a.dest not in {'config', 'verbose', 'version'}) and | ||
| # only use config file value if the value was not explicitly given on command line | ||
| (not given_args.__contains__(a.dest)) |
There was a problem hiding this comment.
| (not given_args.__contains__(a.dest)) | |
| (a.dist not in given_args) |
| @@ -220,6 +220,8 @@ def _parse_and_format_args(): | |||
| """ | |||
| parser = _make_arg_parser() | |||
| args = parser.parse_args() | |||
| # parse arguments without setting defaults | |||
| given_args, foo = parser._parse_known_args(sys.argv[1:], argparse.Namespace()) | |||
There was a problem hiding this comment.
| given_args, foo = parser._parse_known_args(sys.argv[1:], argparse.Namespace()) | |
| given_args, _ = parser._parse_known_args(sys.argv[1:], argparse.Namespace()) |
|
Change of plans: I'll merge this and fix the two style nits in the release PR. |
… while keeping ability to override from command line.
With this change config.yaml can contain most command line parameters and you can have a config file like
If parameters are both found in the config file and in the given command line arguments, the latter will take precedence.
Now you can simply call conda-mirror with --config