Bug fixes #55
Bug fixes #55
Conversation
|
@soniamitchell - can you try running this to make sure you can still import ( |
|
I need this merged with dev to test run --ci |
fair/session.py
Outdated
| # Populate file type table | ||
| _local_uri = fdp_conf.get_local_uri() | ||
| _local_uri = self._global_config['registries']['local']['uri'] |
There was a problem hiding this comment.
Can I ask why have you changed this line? The function get_local_uri does this anyway, but also catches the exception to print it in the same style as all errors in the CLI (hides the traceback unless --debug is used) and ensures it can be called from anywhere. Using a global function is better than hardcoding as it means should the layout of things change, only one place has to be altered.
def get_local_uri() -> str:
_cfg = read_global_fdpconfig()
try:
return _cfg['registries']['local']['uri']
except KeyError:
raise fdp_exc.CLIConfigurationError(
f"Expected key 'registries:local:uri' in local CLI configuration"
)There was a problem hiding this comment.
It doesn't do the same actually - I suspect because get_local_uri() relies on read_global_fdpconfig() which relies on the file being written rather than being in memory?
There was a problem hiding this comment.
@kzscisoft - I totally agree this is inelegant btw, happy to have a fix, but I just don't think it can involve get_local_uri()...
There was a problem hiding this comment.
It doesn't do the same actually - I suspect because
get_local_uri()relies onread_global_fdpconfig()which relies on the file being written rather than being in memory?
Hmm but is written a couple of lines above:
with open(fdp_com.global_fdpconfig(), "w") as f:
yaml.dump(self._global_config, f)Is this write failing?
There was a problem hiding this comment.
It may seem a bit weird to re-read the same thing, but it is a failsafe in a way in that it tests the write actually did happen.
There was a problem hiding this comment.
Hah. Well, that's weird. It works now. I suspect that the file wasn't getting written out correctly because it was looking for a description key somewhere, and I got rid of that requirement - when I looked the file existed but was empty, so I assumed that it was part way through writing for some reason. Anyway, I'll happily undo this change then.
|
Kudos, SonarCloud Quality Gate passed!
|
Bugfixes:
--globwith--globalforfair purge, closesfair purgetakes a --glob argument when it should be --global #35descriptionkey incli-config.yamlfiles sofair init --export <file.yaml>doesn't work, closesfair init --export <file.yaml>no longer works #54fair init --using <file.yaml>, closesfair init --using <file.yaml>no longer works #56