> On Sept. 3, 2013, 4:35 p.m., David Faure wrote:
> > It's now kioslave/file/file.cpp, please adjust indeed.
> > 
> > Don't simplify isReadable: there's 000 (the first if), 444 (the second if), 
> > and other combinations (004, 040, 400) (the case where neither of these 
> > simple tests pass so we must use additional API to check for readability).
> > You suggestion "a 4 anywhere" doesn't make the file necessarily readable to 
> > us (if we don't own it).

But then the second if should be "(S_IRUSR|S_IRGRP|S_IROTH) == 
d->m_permissions", not "(S_IRUSR|S_IRGRP|S_IROTH) & d->m_permissions", no?


- Aurélien


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/112481/#review39266
-----------------------------------------------------------


On Sept. 3, 2013, 4:04 p.m., Aurélien Gâteau wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/112481/
> -----------------------------------------------------------
> 
> (Updated Sept. 3, 2013, 4:04 p.m.)
> 
> 
> Review request for KDE Frameworks and David Faure.
> 
> 
> Description
> -------
> 
> Partly revert commit b03e81a61311ae1b64b0d37415477f9c08fe6142 which ported 
> KFileItem to QFile::Permissions.
> 
> Quoting David Faure:
> 
> "Port back from QFile::Permissions to mode_t given that mode_t is available in
> win32-msvc*/qplatformdefs.h (not everywhere, but in the places where the above
> commits did a half job at porting from mode_t to QFile::Permissions).
> 
> QFile::Permissions is more limited (no suid bits) and having to convert 
> between
> mode_t and QFile::Permissions is just extra work."
> 
> I am not sure about the following part of KFileItemPrivate::init():
> 
>     } else { // link pointing to nowhere (see kio/file/file.cc)
>         mode = (QT_STAT_MASK - 1) | S_IRWXU | S_IRWXG | S_IRWXO;
>     }
> 
> I blindly brought back from the old code without understanding it. If 
> anything, the comment needs to be updated as there is no kio/file/file.cc 
> anymore.
> 
> I am also wondering if this part of the code in isReadable() could be 
> simplified a bit:
> 
>     // No read permission at all
>     if (!(S_IRUSR & d->m_permissions) && !(S_IRGRP & d->m_permissions) && 
> !(S_IROTH & d->m_permissions))
>         return false;
> 
>     // Read permissions for all: save a stat call
>     if ((S_IRUSR|S_IRGRP|S_IROTH) & d->m_permissions)
>         return true;
>  
> Isn't this the same as this?
> 
>     return (S_IRUSR|S_IRGRP|S_IROTH) & d->m_permissions;
> 
> Same thing for isWritable().
> 
> 
> Diffs
> -----
> 
>   kio/tests/kfileitemtest.cpp c8b7a5d 
>   staging/kio/src/core/kfileitem.h 623b426 
>   staging/kio/src/core/kfileitem.cpp 970dea8 
> 
> Diff: http://git.reviewboard.kde.org/r/112481/diff/
> 
> 
> Testing
> -------
> 
> All tests pass.
> 
> 
> Thanks,
> 
> Aurélien Gâteau
> 
>

_______________________________________________
Kde-frameworks-devel mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel

Reply via email to