When content is processed and found to be valid JSON it is decoded into a dict. To write it out to a file we need to encode it back into a string.
Addresses GH-5914.
It turns out that some of the assumptions in #5885 were slightly off. The previous fix relied on a call to the module to creat a tmp_path. This is insufficent as there are few cases that we need to have the tmp directory before we make the module call. If we don't have a tmp_path before we do a recursive call or when we find a file that does not match the remote md5 hash we need to create a tmp directory. Also we are not more percise when we will need to clean up the remote tmp_path.
The copy action_plugin is not easy to read. Part of this commit is taking that file, restructuring it, and adding comments. No functionality changed in how it interacts with the world.
The fix for #5739 ends up being the assumption that there is a cleanup 'rm -rf' that happens at the end of the copy loop. This was not the fact before and we made a bunch of tmp directories that we hoped would end up being cleaned up. Now we just use the tmp directory that the runner provides and cleanup inline if it is a single file to be coppied or after the loop if it is a recursive copy.
As a part of this we did end up having to change runner to provide a flag so that we could short the inline tmp directory removal. This flag defaults to True so it will not change the behavior of other modules that are being called.
The copy action plugin creates its own tmp dir for each file that it
copies to the target machine. However, it does not clean up the
original tmp path it was given when run() is called. This cleans up the
tmp path before it begins looping on source files.
For this, add internal "original_basename" param to file module,
similar to copy module. (Param name is a bit misnormer now, should
be treated as "original basepath").
If src param to copy is a directory, all files under it are collected
and pushed one by one to target. Source dir path handled in a way
simalar to rsync: if it ends with slash, only inside contents of
directory are copied to destination, otherwise the dir itself is
copied (with all contents of course). Original idea and implementation
by https://github.com/ansible/ansible/pull/1809 . Rewritten to address
review comments and simplify/correct logic.
The 'always_run' task clause allows one to execute a task even in
check mode.
While here implement Runner.noop_on_check() to check if a runner
really should execute its task, with respect to check mode option
and 'always_run' clause.
Also add the optional 'jinja2' argument to check_conditional() :
it allows to give this function a jinja2 expression without exposing
the 'jinja2_compare' implementation mechanism.
Previously setting force=no caused copy to subversively
fail when target did not exist on remote host.
Caused by Runner._remote_md5 returning 1
when files don't exist, rather than 0.
The copy action accepts force=no, which tells it not to replace an
existing file even if it differs from the source. The copy action
plug-in wasn't respecting this option when operated in check mode, so it
would report that changes are necessary in check mode even though copy
would make no changes when run normally.
Runner._remote_md5 was changed to make the logic for setting rc perhaps
a little more clear, and to make sure that rc=0 when the file does not
exist.
If it is a directory, change the destination path by appending the
basename of the source file, like is done if the destination ends with a
/, and try to get the MD5 of the new path.
