oranenj

the existing acceptance tests do test a part of the new providers' functionality. In the fcontext case, I extended the test to even check if purging works. As of this...

I'm testing with a branch where all three providers have been merged together. Currently successful with CentOS 7.3, and I think I've fixed the remaining issues with CentOS6 as well....

So acceptance tests exist, but they might not be complete as complete as they should be, so I'm still keeping this open

This seems like a bit of a hack, but I'm not opposed to the idea. You might need to call matchpathcon_init again though, maybe? Reading the manpage of the matchpathcon...

@ekohl do you know if Puppet uses multiple threads when applying manifests? Looking at the matchpathcon source code, it seems to use thread-local storage in some manner that might affect...

This might be quite non-trivial to implement. essentially it seems we'd have to reimplement the selmodule type entirely... Which might happen eventually, anyway, considering how buggy it currently is. Additionally,...

I suppose the specs could use a look-over, but the purpose of the before/after all blocks is to set up some state in the system for the module to act...

@cschanzlenist as far as I can tell, the code has nothing that removes externally managed *base* repositories from the configuration, it seems to merely assume that removing the repos package...

@cschanzlenist I don't really control which file gets edited, since the configuration is enforced by Puppet, and it'll parse the repo files, remove any repositories that are not managed, and...

Something like this seems to fix this ``` +++ b/sway/tree/container.c @@ -1705,10 +1705,12 @@ static bool is_parallel(enum sway_container_layout first, static bool container_is_squashable(struct sway_container *con, struct sway_container *child) { enum sway_container_layout...