D8387: Workaround incorrectly returned EEXIST instead of EPERM regression introduced by libsmbclient 4.7

Michal MalĂ˝ noreply at phabricator.kde.org
Sat Oct 21 11:09:04 UTC 2017


madcatx added inline comments.

INLINE COMMENTS

> elvisangelaccio wrote in kio_smb_browse.cpp:521
> Don't we set `m_enableEEXISTWorkaround` in the constructor?

Yes. The idea is this: When the slave loads, it checks what libsmbclient lib is available and sets the `m_enableEEXISTWorkaround` if a broken version is detected. Then iff the return code from `smbc_opendir()` is EEXIST and the `m_enableEEXISTWorkaround` is true EEXIST is treated as EACCES or EPERM.  This should hopefully prevent any weird behavior when EEXIST might be a valid return code. Do you still think that the method that does this "apply the workaround" logic could use a better name?

REPOSITORY
  R320 KIO Extras

REVISION DETAIL
  https://phabricator.kde.org/D8387

To: madcatx, ngraham, davidedmundson, elvisangelaccio, #frameworks
Cc: graesslin, z3ntu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20171021/5a5c4612/attachment.html>


More information about the Kde-frameworks-devel mailing list