Another case sensitivity issue (symlink and rename involved, of course - MacOS only)

[agroce]

import pyfakefs.fake_filesystem
import os

if os.path.exists('/Volumes/ramdisk/test'):
shutil.rmtree('/Volumes/ramdisk/test')
os.mkdir('/Volumes/ramdisk/test')

fs = pyfakefs.fake_filesystem.FakeFilesystem()
fs.CreateDirectory('/Volumes')
fs.CreateDirectory('/Volumes/ramdisk')
fs.CreateDirectory('/Volumes/ramdisk/test')
os0 = pyfakefs.fake_filesystem.FakeOsModule(fs)

component0 = "beta"
component1 = "b"
path0 = "/Volumes/ramdisk/test"
path0 += "/" + component0
component0 = "Beta"
path1 = "/Volumes/ramdisk/test"
path2 = "/Volumes/ramdisk/test"
os0.mkdir(path0)
path0 = "/Volumes/ramdisk/test"
path0 += "/" + component1
os0.symlink(path1,path0)
path0 += "/" + component0
path1 += "/" + component0
os0.rename(path0,path1)
result = sorted(os0.listdir(path2))

print result

pyfakefs: ['Beta', 'b']
os: ['b', 'beta']

Thought we were done, this one took some actual CPU time to generate. Testing at depth 100 didn't find anything, 200 took nearly 5 minutes to hit this, which is really a while.

[mrbean-bremen]

Yes, I'm not surprised. As I said, I have no idea what the rules are under MacOS (as opposed to Windows, where it seems straightforward to me). So I'm running pure TDD here - just ensuring that the test passes and waiting for the next failing test...

[agroce]

I'm not surprised in the process, but TSTL taking that long running without coverage is always a bit of a shock. We basically ran this "TDD" process the same way for the JPL file systems, just closing the loop slightly quicker since the deadlines were much closer and it was everyone's day job (and IMO the file systems were a lot simpler, in some ways).

[agroce]

I suspect once I dig into more mutants, we'll have several more bursts of things uncovered like this.

[mrbean-bremen]

Yes, I thought about checking the mutants myself, but didn't get to it. I wanted to fix the bugs I found in the pyfakefs tests vs. real FS first (issue #303) - which are mostly problems under Windows - but this has been stalled because of a Windows problem I'm having on my notebook after the latest Windows 10 update. Actually not the first problem - it just got worse after the last update. Sigh...
Until I resolve that, I will work under Linux, which is ok, but not good to find the Windows specific stuff, and also the stuff specific for case-insensitive fs. Maybe I will setup AppVeyor on GitHub to be able to run the Windows tests automatically.
Anyway, I have also tried to fix the current issue and got stuck because I really don't understand the logic behind. I don't have much free time at the moment, so this may take a while... Maybe with the next mutants you check there will come up some easy-to-fix issues that I can tackle meanwhile.