Add stat times#13
Conversation
|
Travis failure is due to the lack of py26 on the travis image, and the decrease in coverage is mostly due to me marking two tests for skipping because I couldn't figure out if they were right or not. This PR merges cleanly with #11, but will need one line modified after the merge in |
WeatherGod
force-pushed
the
add_stat_times
branch
from
42edf2a to
c8846bd
Compare
|
I rebased this PR onto master. |
|
Interesting that this fails on windows only, but not linux. The failures on windows basically show that the mtimes aren't getting updated. |
| return len(self.get(path, atime_change=False)) | ||
| except TypeError: | ||
| return len(str(self.get(path))) | ||
| return len(str(self.get(path, atime_change=False))) |
There was a problem hiding this comment.
Just noticed that this would be wrong in py3k in light of #15 (which doesn't fix this).
ulope
force-pushed
the
master
branch
13 times, most recently
from
bfaaa6d to
95a3bc8
Compare
|
Let me know if you'd like me to try rebasing this branch to take into account all of the recent changes. I did it once before, so maybe I can find that branch somewhere... |
|
@WeatherGod if you could that would be nice! |
* added recursive_list() method * parametrized some existing tests * marked a broken test to skip * added tests to ensure times are managed correctly before and after file operations
it will trigger a rebuild of the time dictionary. This is needed for the .serve_content() method of sftpserver. Also refactor the tests a bit to reduce redundant code.
time-handling features of ContentProvider. Also refactor the unit tests to reduce redundant code.
WeatherGod
force-pushed
the
add_stat_times
branch
from
c8846bd to
b2c69ce
Compare
|
rebased and forced-pushed. Also addressed the comment I made about the additional line of code that needed merging. Don't forget that my changes also includes marking two existing tests for skipping, as I think they are wrong, but not being the original author, I can't be certain. |
|
seriously, a simple, small, recursive function is considered "too complex", and for a concept that nearly any programmer should understand (recursive listing of a directory structure...)? Bah! |
|
Thanks!
I’ll have a look. As for the code climate warnings, I agree threat they are a bit over the top. The thresholds seem to be configurable. I’ll see if I can chill it out a bit. But it’s not a mandatory check anyway. |
|
Weird, this is still showing the same problem on Windows that I noticed before where the mtimes aren't getting updated. This problem doesn't happen on my linux systems. I am not sure why this would only impact windows, unless maybe paramiko is doing something different on windows than on linux? I am not familiar enough with windows to know. |
3 participants
sftpserver should now manage atime and mtime of the "files", and operating on the filesystem should get you proper atime/mtime modification behavior.
Needed for writing tests for an internal package of my employer that checks the lateness of product deliveries.
I had to mark two old test to skip because they no longer work, but I am not convinced the tests were correct before.