The gitman install command will now properly recurse through folders. It will ignore any nested dependencies within currently installed gitman repositories.
The filter_nested_configs function will now properly remove any nested configs from a list of configs. This does a check between all configs to make sure it is not belonging to any other configs.
Removed settings.CI from overriding find_nested_configs as it was breaking our and this CI from running without unsetting the variable.
In fixing the CI bug, I also had to update the make demo command. Now they all explicitly call gitman <command> --depth 1. I am not sure how else to fix this other than running these commands in a subdirectory that only contains the one gitman.yml file.
I added depth to gitman lock which was already an unexposed argument. This was to appease make demo. Ideally I would also expose the depth argument to uninstall but this PR is already getting large enough.
I have also added a test to assert that the changes in this PR are functional in the future.
In regards to issue https://github.com/jacebrowning/gitman/issues/311, this PR addresses the five following issues:
gitman install
command will now properly recurse through folders. It will ignore any nested dependencies within currently installedgitman
repositories.filter_nested_configs
function will now properly remove any nested configs from a list of configs. This does a check between all configs to make sure it is not belonging to any other configs.settings.CI
from overridingfind_nested_configs
as it was breaking our and this CI from running without unsetting the variable.CI
bug, I also had to update themake demo
command. Now they all explicitly callgitman <command> --depth 1
. I am not sure how else to fix this other than running these commands in a subdirectory that only contains the onegitman.yml
file.depth
togitman lock
which was already an unexposed argument. This was to appeasemake demo
. Ideally I would also expose thedepth
argument touninstall
but this PR is already getting large enough.I have also added a test to assert that the changes in this PR are functional in the future.
Closes #311