[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Tramp misunderstanding AFS ACLs
From: |
Michael Albinus |
Subject: |
Re: Tramp misunderstanding AFS ACLs |
Date: |
Thu, 16 Jul 2009 22:56:19 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/23.1.50 (gnu/linux) |
Steven Hirsch <address@hidden> writes:
>> Anyway, one would need a mapping of AFS ACLs into file permission bits,
>> because that's what `file-attributes' asks for. Returning stupid
>> "-rwxrwxrwx" does not look like a clever solution; I see the wave of bug
>> reports rolling on.
>
> I agree that it should not be a default, but why not have a "dare to
> be stupid" toggle if you really want to live dangerously?
>
> Mapping AFS ACLs to Unix permissions is not particularly trivial, and
> would require that the AFS 'fs' command be available on the path at
> the remote side.
>
> How do you obtain the Unix bits now? By parsing the output from 'ls'?
> Maybe this could be done by writing a little shell script shim to the
> far end and parsing in the output.
There are three functions: `tramp-handle-file-attributes-with-stat',
`tramp-handle-file-attributes-with-perl', and
`tramp-handle-file-attributes-with-ls'. They are tried in this order,
see `tramp-handle-file-attributes'.
One could massage file permissions in `tramp-convert-file-attributes',
which is called finally in all three cases. This could depend on a new
option `tramp-please-adapt-my-file-permissions-to-be-stupid', or so.
But there are also the functions `tramp-handle-file-readable-p' and
`tramp-handle-file-writable-p', which call their own test functions,
IIRC.
If you have a patch for this, you might show it on the list.
> Steve
Best regards, Michael.