https://bugs.kde.org/show_bug.cgi?id=421228
Sebastian Hirsch <sebastian_hir...@posteo.de> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|UPSTREAM |FIXED
--- Comment #19 from Sebastian Hirsch <sebastian_hir...@posteo.de> ---
(In reply to Harald Sitter from comment #18)
> I think I have all the info I need, thanks.
>
> That is one crazy issue though...
>
> What happens is that we list the directory contents and it gets correctly
> detected as a regular file
>
> > log_kio_smb: [ "UDS_NAME"="Manual_TX-8020_ItDeNlSv.pdf"
> > "UDS_FILE_TYPE"=32768 "UDS_SIZE"=22967500
> > "UDS_MODIFICATION_TIME"=1576488632 "UDS_ACCESS_TIME"=1576488632 ]
>
> after that we do get details on the directory for one reason or another,
> which as one would expect is reporting it is a directory
>
> > log_kio_smb: [ "UDS_NAME"="." "UDS_FILE_TYPE"=16384 "UDS_SIZE"=0
> > "UDS_MODIFICATION_TIME"=1586423295 "UDS_ACCESS_TIME"=1588930044 ]
>
> we then get details on the file and now the file suddenly is a directory :O
>
> > log_kio_smb: [ "UDS_NAME"="Manual_TX-8020_ItDeNlSv.pdf"
> > "UDS_FILE_TYPE"=16384 "UDS_SIZE"=0 "UDS_MODIFICATION_TIME"=1586423295
> > "UDS_ACCESS_TIME"=1588930044 ]
>
> What I suspect happens is that libsmbclient internally has a caching problem
> of some sort. The data we get on the file the second time around looks
> suspiciously similar to the data of the directory (mtime and atime are the
> same).
> That theory is further supported by it suddenly getting reported as a file
> again on a request shortly afterwards.
>
> I've filed a bug report on samba about this.
> https://bugzilla.samba.org/show_bug.cgi?id=14379
Thanks a lot for your investigation! Lets see if this issue gets fixed.
--
You are receiving this mail because:
You are watching all bug changes.
