El jue, 16-06-2005 a las 10:50 +0200, Werner Baumann escribi=F3:
> my suggestions concerning the bugs in bugs.debian.org:
> #303533 CODA kernel version mismatch:
> Because of different Coda interfaces I think it's hard to make one=20
> davfs2-package working with both kernel versions. As I think davfs2 is=20
> mainly for desktops and kernel 2.4 will vanish on desktops soon, I=20
> dont't no whether it is worth the effort. So a transitional solution=20
> with two conflicting versions would be the easy way (at least for me if=20
> Luciano does the work ;-) ).
I saw you suggestion and it's look very interesting. The portability
of a program is something something incredibly valuable. And it's less
confusing to the users. It's the better solution, but I don't know how
difficult is implemented it.
> #311286 WebDAV URL's with spaces do not work:
> This is a matter of shell escaping. I have send a reply explaining to=20
> escape spaces in arguments and options with '\' and not to do any http=20
> encoding, as this will be done by davfs2.
This bug is marked as done. I will add an example with spaces in the man
> #310976 convert usernames/groupsnames to uid/gid:
> I think this is solved in davfs2-0.2.3-testing. But it is disabled at=20
> the moment because of my quick an brutal security fix. For security=20
> reasons user and group are fixed to the mounting user.
> #307096 parse environment for password, username and proxy:
> davfs2-0.2.3-testing (cvs) will take the proxy from environment if not=20
> given another way.
> But I don't like the idea of taking username and password from=20
> environment too. If a WebDAV-resource is used regularly the=20
> creditentials should be in the secrets file; else they should be asked=20
> interactively. This is the way davfs-0.2.3-testing does it.
> #310757 doesn't enforce permissions:
> I have done a fix for this. This is not the ultimate solution but I=20
> think a clean solution should be combined with integration of a=20
> directory cache (I'm working on this).
> Unfortunately my fix for davfs2-0.2.3 does no longer allow user mount as=20
> this would be a security hole.
> davfs2-0.2.3-testing has almost the same patch but allows user mount as=20
> user mount is handled differnt to davfs2-0.2.3.
The patch ignore the --mode option. The testing branch resolve this?
> I think, except #303533 CODA kernel version mismatch, all bugs are=20
> solved in davfs2-0.2.3-testing. So please test.
> Would you mind if I merged davfs2-0.2.3-testing back into the main=20
> branch? I now think this might be a base for the next version (mainly=20
> because of improved user mount and proper behaviour on unmount). But we=20
> need to find the bugs that I don't see. Please tell me.
> I started working on a directory cache as I think there are far to many=20
> request to the WebDAV-server that slow down davfs2 severely. So I would=20
> like to use the 0.2.3-testing branch for this. But as this will take=20
> some time to work I don't want to mix it up with the present=20
> davfs2-0.2.3-testing version.
For months you have been the most active developer. I believe that,
unless somebody complains, you have freedom to do what it seems to you
the best thing. This is just my point of view.
Luciano Bello <luciano@...>