Tested on ZFS filesystem


    [test0@s0]:<~>$ mkdir 
'232222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222'
    [test0@s0]:<~>$ cd 
232222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222/
    
[test0@s0]:<~/232222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222>$
 git status
    fatal: Unable to read current working directory: Permission denied

    
[test0@s0]:<~/232222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222>$
 git blahblahblah
    fatal: Unable to read current working directory: Permission denied


But when I create directory with same name (232...) as it's subdirectory 
everything works as expected inside that subdirectory.Od: "René Scharfe" 
&lt;l....@web.de>
Do: "Zenobiusz Kunegunda" &lt;zenobiusz.kunegu...@interia.pl>; 
Wysłane: 22:18 Sobota 2017-03-11
Temat: Re: fatal: Could not get current working directory: Permission denied | 
affected 2.10,2.11,2.12, but not 1.9.5 |

> 
>> Am 09.03.2017 um 12:01 schrieb Zenobiusz Kunegunda:
>> OK, I printed euids inside function calling getcwd(). Every single
>> EUID is the same and have expected value. The same as any other local
>> application run by this user. Permissions of every directory in the
>> path are OK.
>>
>> /bin/pwd -P inside directory returned exact path and exited with exit
>> status 0.
> 
> Weird.
> 
> I tried to reproduce the behavior on FreeBSD 10.3, unsuccessfully.  I 
> tried with Ruby 2.2 from the ports tree and Ruby 2.3 installed with 
> rbenv, and with git 2.6.4 from the ports tree and a self-compiled 
> version of the current master branch.
> 
> You could try to bisect the issue to find which commit to git caused the 
> behavior change, but that will be a tedious process (compile and install 
> an intermediate version, run bundle install up to the point of interest 
> then hit Ctrl-C, clean up somehow, run "git bisect good" or "git bisect 
> bad" in the git source tree depending on the outcome of bundle, repeat 
> ca. twelve times more).
> 
> But I suspect the cause of the problem hides somewhere else.  Why would 
> pwd(1) be able to call getcwd(3) just fine, while git gets an EPERM 
> error for the same call?  I'm stumped.
> 
> René
> 
> 

Reply via email to