> In general, the /tmp directory on most Unix systems (including Linux)
> should
> be cleared at reboot. I am surprised to hear that there are systems that
> don't do that.

For all I can say, Suse doesn't.

> Even if a reboot does occur only rarely (Unix servers often aren't
> rebooted
> for months if not years, as a reboot is required only rarely compared to
> other OSes), the cache data might still be wanted to survive (and I don't
> really see a reason why they shouldn't).

Agreed.

>> so we don't have to worry about it. Since we are running
> ...
> On the other hand, as Derrell points out, it might not be the ideal
> location
> IF the data in this directory is supposed to survive a reboot.

Understood. BTW, I didn't finished the sentence starting "Since we are
running...", but it was about running on three OS's, so we are looking for
a cross-platform location (but everybody probably guessed that).

> It seems to make sense to me to configure the cache location in the
> config.json file on a per application basis. This way each developer can
> decide what he wants to do.

Well, as I said this is always possible. Burak wanted it to be explicit in
the config.json. Do you think the same, or are you satisfied with an
implicit default - explicit deviation?

> Personally, I think it would make sense to have the cache directory
> default
> to a directory inside the application tree.

This was exactly what we wanted to avoid. I can live with a cache location
outside /tmp, eg. in the qooxdoo path, but I would want to retain the
cross-application location, so you have the benefit of cache reuse across
applications. You can always change that in the individual config.json's,
ok?!

> Perhaps /var/tmp/ would be more reasonable, although I am not sure if
> there
> aren't distributions that clear this as well (Ubuntu Jaunty doesn't).

Yeah, I wouldn't want to rely on that too.

>> As I said, the solution is as close as your config.json. As for the poor
>> user experience due to automatically erased /tmp directories, I would be
>> interested to learn who else has this issue?!
>
> Just tried my current application with
>
> ./generate source
>
> With cache:    1.592u 0.112s 0:01.76 96.5%    0+0k 1944+0io 0pf+0w
> Without cache:  56.003u 0.956s 1:08.43 83.2%    0+0k 48432+0io 0pf+0w
>
> So this WILL hurt, if the cache is deleted unintentionally.

Thanks for the metrics, Fritz, but I was not questioning the difference
between cache and no cache, I was asking about the /tmp clearance :).

T.



------------------------------------------------------------------------------
OpenSolaris 2009.06 is a cutting edge operating system for enterprises 
looking to deploy the next generation of Solaris that includes the latest 
innovations from Sun and the OpenSource community. Download a copy and 
enjoy capabilities such as Networking, Storage and Virtualization. 
Go to: http://p.sf.net/sfu/opensolaris-get
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to