postfixadmin-devel

[Postfixadmin-devel] big merge/commit of smarty branch into trunk

[reg9009 <re...@ya...>]

Hi all,

I've merged the actual smarty branch into trunk. The branch was
up-to-date with all recent changes made.

The only important thing is that the directory "/templates_c" needs to
be writable to the web server process, as compiled templates are stored
in this directory.

Regards,
Sebastian

Re: [Postfixadmin-devel] big merge/commit of smarty branch into trunk

[David G. <da...@co...>]

reg9009 wrote :
> Hi all,
> 
> I've merged the actual smarty branch into trunk. The branch was
> up-to-date with all recent changes made.
> 
> The only important thing is that the directory "/templates_c" needs to
> be writable to the web server process, as compiled templates are stored
> in this directory.
> 

OK.

How about....

a) setup.php has a check to see whether it exists and is writeable

or

b) if templates_c is not writeable, we use e.g. /tmp/?

(Thanks for merging it)

David.
-- 
David Goodwin 

[ david at codepoets dot co dot uk ]
[ http://www.codepoets.co.uk       ]

Re: [Postfixadmin-devel] big merge/commit of smarty branch into trunk

[reg9009 <re...@ya...>]

> OK.
>
> How about....
>
> a) setup.php has a check to see whether it exists and is writeable
>
> or
>
> b) if templates_c is not writeable, we use e.g. /tmp/?
>
> (Thanks for merging it)
>
> David.
>   

Sounds good, will try to improve error handling...

Regards,
Sebastian

Re: [Postfixadmin-devel] big merge/commit of smarty branch into trunk

[Christian B. <pos...@cb...>]

Hello,

Am Donnerstag, 5. November 2009 schrieb David Goodwin:
[templates_c]
> a) setup.php has a check to see whether it exists and is writeable

Good idea. Just try to write some text to a test file ;-)

> b) if templates_c is not writeable, 

Then display an error message ;-)

> we use e.g. /tmp/?

Please don't do this. The filenames in templates_c are predictable and 
therefore at the risk of symlink attacks.

BTW: Is there a way to make smarty working without the templates_c 
directory? If yes, it could serve as a fallback solution and the error 
message could be "just" a warning. (The reduced performance because of 
the missing cache would be better than blocking everything ;-)


Regards,

Christian Boltz
-- 
Ossi-Virus
Güdn tog, isch bün än Häggr aüs Läibzsch un dis is än selbsdprögammirdr
bösardscher Gommbjudrfirus. Do isch nö nit sö vül wäis von döm Gommbjudr
is dis än mänuoöllr Firus. Alsö löschnse bidde ölle Dodein fön de
Fesdpladde und schiggnse disn Firus ön ölle di se genn.
- Dangge für de Midorbeid -

Re: [Postfixadmin-devel] big merge/commit of smarty branch into?trunk

[David G. <da...@co...>]

Christian Boltz wrote :
> Hello,
> 
> Am Donnerstag, 5. November 2009 schrieb David Goodwin:
> [templates_c]
> > a) setup.php has a check to see whether it exists and is writeable
> 
> Good idea. Just try to write some text to a test file ;-)
> 
> > b) if templates_c is not writeable, 
> 
> Then display an error message ;-)
> 
> > we use e.g. /tmp/?
> 
> Please don't do this. The filenames in templates_c are predictable and 
> therefore at the risk of symlink attacks.
> 
> BTW: Is there a way to make smarty working without the templates_c 
> directory? If yes, it could serve as a fallback solution and the error 
> message could be "just" a warning. (The reduced performance because of 
> the missing cache would be better than blocking everything ;-)
> 

Yes; you can tell it to not compile templates (i.e. re-compile every
template every time etc). No one is likely to notice the performance
hit of this.

David.

-- 
David Goodwin 

[ david at codepoets dot co dot uk ]
[ http://www.codepoets.co.uk       ]

Re: [Postfixadmin-devel] big merge/commit of smarty branch into?trunk

[Christian B. <pos...@cb...>]

Hello,

Am Freitag, 20. November 2009 schrieb David Goodwin:
> Christian Boltz wrote :
> > BTW: Is there a way to make smarty working without the templates_c
> > directory? If yes, it could serve as a fallback solution and the
> > error message could be "just" a warning. (The reduced performance
> > because of the missing cache would be better than blocking
> > everything ;-)
> 
> Yes; you can tell it to not compile templates (i.e. re-compile every
> template every time etc). No one is likely to notice the performance
> hit of this.

Sounds interesting :-)
Can you implement this, please?

I'm not sure if disabling templates_c in general is the best idea, but 
having a fallback if templates_c isn't writeable or doesn't exist would 
be good.


Regards,

Christian Boltz
-- 
> You only read the second paragraph, didn't you?
Why do you write emails where one has to read the stuff
between the first and the last word?
[> Stephan Kulow and Dirk Mueller in opensuse-packaging]