dovecot - Mar 2015 - Why is Sieve trying to re-compile global scripts?

On 3/12/2015 11:56 AM, Olaf Hopp wrote:
> On 03/12/2015 12:02 AM, Stephan Bosch wrote:
>> On 3/11/2015 11:10 AM, Olaf Hopp wrote:
>>> Please see the thread with subject
>>> "Sieve permissions issue following update"
>>> I tested sucessfully a developper issue last month
>>> on the hint of Stephan. Yesterday I started to test the currenr
RCs.
>>>
>>> First I was disappointed, because the error seems to persist.
>>> So I double checked everything, recreated / recompiled everything
>>> an the error went away. So I thought it was mistake on my side.
>>> I gave Spephan postive feedback. And I'm waiting for the final
release
>>> for my production server.
>>>
>>> But when I read your mails, I'm not feeling happy.
>>> I think it's a kink of luck/voodoo/whatever.
>>>
>>> What you must do, I think, is to compile the sieve script with the
>>> exact version running afterwards.
>>> And I think you should the remove the compiled .svbin files
>>> before recreating them again. Don't overwrite them with the
compiler.
>>>
>>> I think I'll also dig into this any further today.
>>
>> Please do. I cannot reproduce this so far.
>>
>> Since E.B. still got an obscure debug message about metadata not being
>> up to date, I added debug lines to the remaining places where this
could
>> emerge (currently only available from hg).
>>
>> Regards,
>>
>> Stephan.
>>
>
> Hi,
> I'm still trying but currently I can not reproduce the bug.
> But I will keep on hammering on it.
Looks like I found the bug. Will need some time to fix this properly.

Regards,

Stephan.

On 3/12/2015 11:53 PM, Stephan Bosch wrote:
> On 3/12/2015 11:56 AM, Olaf Hopp wrote:
>> On 03/12/2015 12:02 AM, Stephan Bosch wrote:
>>>
>>> Please do. I cannot reproduce this so far.
>>>
>>> Since E.B. still got an obscure debug message about metadata not
being
>>> up to date, I added debug lines to the remaining places where this
could
>>> emerge (currently only available from hg).
>>>
>>> Regards,
>>>
>>> Stephan.
>>>
>> Hi,
>> I'm still trying but currently I can not reproduce the bug.
>> But I will keep on hammering on it.
> Looks like I found the bug. Will need some time to fix this properly.
I released rc2. Please check whether this resolves the issues.

Regards,

Stephan.

On 03/15/2015 12:37 AM, Stephan Bosch wrote:
> On 3/12/2015 11:53 PM, Stephan Bosch wrote:
>> On 3/12/2015 11:56 AM, Olaf Hopp wrote:
>>> On 03/12/2015 12:02 AM, Stephan Bosch wrote:
>>>>
>>>> Please do. I cannot reproduce this so far.
>>>>
>>>> Since E.B. still got an obscure debug message about metadata
not being
>>>> up to date, I added debug lines to the remaining places where
this could
>>>> emerge (currently only available from hg).
>>>>
>>>> Regards,
>>>>
>>>> Stephan.
>>>>
>>> Hi,
>>> I'm still trying but currently I can not reproduce the bug.
>>> But I will keep on hammering on it.
>> Looks like I found the bug. Will need some time to fix this properly.
>
> I released rc2. Please check whether this resolves the issues.
>
With RC2 everything looks good !

And finally I could reproduce the bug:
with 0.4.5 and 0.4.7 RC1 you can trigger it when you compile
the master sieve script with a *relative* path:

cd /etc/dovecot
/usr/bin/sievec -D ./sieve-master

will trigger it. Whereas
      /usr/bin/sievec -D /etc/dovecot/sieve-master
even with 0.4.5 will run fine.

With 0.4.7 RC2 it makes no difference, wether you use an absolute
or a relative path to the sieve-master script.

Olaf

-- 
Karlsruher Institut f?r Technologie (KIT)
ATIS - Abt. Technische Infrastruktur, Fakult?t f?r Informatik

Dipl.-Geophys. Olaf Hopp
- Leitung IT-Dienste -

Am Fasanengarten 5, Geb?ude 50.34, Raum 009
76131 Karlsruhe
Telefon: +49 721 608-43973
Fax: +49 721 608-46699
E-Mail: Olaf.Hopp at kit.edu
www.atis.informatik.kit.edu

www.kit.edu

KIT - Universit?t des Landes Baden-W?rttemberg und nationales Forschungszentrum
in der Helmholtz-Gemeinschaft

Das KIT ist seit 2010 als familiengerechte Hochschule zertifiziert.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5214 bytes
Desc: S/MIME Cryptographic Signature
URL:
<http://dovecot.org/pipermail/dovecot/attachments/20150316/2371ee7b/attachment.p7s>