v2.3.3 rc1 - Error: sieve: !!BUG!!: Binary compiled from dovecot.sieve is still corrupt

[Stephan Bosch]

Op 27-9-2018 om 11:42 schreef James:
> On 25/09/2018 22:39, Stephan Bosch wrote:
>
>> Something mightily weird is going on at your end. It doesn't fail here
>> (see below). First of all, what is your configuration (output from
>> `dovecot -n`)?
>
> You have dovecot.conf but...
>
> $ dovecot -c dovecot.conf -n
> # 2.3.3.rc1 (14e4920d8): dovecot.conf
> # Pigeonhole version 0.5.2 (7704de5e)
> # OS: SunOS 5.11 i86pc
> # Hostname: mailhost
> doveconf: Warning: please set ssl_dh=</etc/opt/xxx/dovecot/dh.pem
> doveconf: Warning: You can generate it with: dd 
> if=/var/opt/xxx/lib/dovecot/ssl-parameters.dat bs=1 skip=88 | openssl 
> dhparam -inform der > /etc/opt/xxx/dovecot/dh.pem
> mail_debug = yes
> mail_gid = staff
> mail_location = maildir:/path/to/%d/%n/Maildir
> mail_uid = james
> managesieve_notify_capability = mailto
> managesieve_sieve_capability = fileinto reject envelope 
> encoded-character vacation subaddress comparator-i;ascii-numeric 
> relational regex imap4flags copy include variables body enotify 
> environment mailbox date index ihave duplicate mime foreverypart 
> extracttext
> postmaster_address = postmaster at domain.tld
> ssl_dh = # hidden, use -P to show it
>
>
>
>> Also, can you make a hex dump of the binary (using `sieve-dump -h
>> <your-script-binary>.svbin`).
>
> As said the svbin is identical to the one create by the previous version.
>
> Comparing the dump:
> Block 0 differs because it has the source file name.
> Block 1 is identical
> Block 2 is identical. It is this block that is declared corrupt.
>
>
>
>> Finally, can you try to explicitly delete the binary (preferably after
>> preserving it elsewhere) so that it is guaranteed to be created fresh?
>
> I did each time.
>
>
>
>> ## Success at my end:
>
> Spot the difference...
>
>> Address   Line  Code
>> 00000000:       DEBUG BLOCK: 3
>> 00000001:       EXTENSIONS [1]:
>> 00000002:         vacation
>> 00000004:    3: VACATION
>> 00000007:    5:   seconds: NUM 86400
>
> Address   Line  Code
> 00000000:       DEBUG BLOCK: 3
> 00000001:       EXTENSIONS [1]:
> 00000002:         vacation
> 00000004:    2: VACATION
> 00000007:    4:   seconds: NUM 5
> 00000009:         Binary is corrupt.
>
> The line numbers differs and 86400 is read as 5.  It is like it has 
> forgotten the size of an integer or is confused about endianness.  
> There is something strange, like an #if that guesses wrong.  At least 
> I have somewhere to start looking.
>
> Thank you for checking at your end, I was worried the RC had 
> introduced an error and your result suggests not.  RCs are for testing 
> and I am.

The number is stored as a chain of bytes of which the most significant 
bit indicates whether the next byte still belongs to the number. If this 
bit is somehow interpreted wrong, the first byte of this number would 
read as 5, thereby returning '5' as the result and ignoring subsequent 
bytes (causing corruption at the next item to read).

Since you're using SunOS, your compiler may be doing something funky. 
Which compiler is used anyway? Perhaps different versions for the 
Dovecot releases that do and don't work?

Does it help when you change the "> 0" at the following code position to 
"!= 0" ?

https://github.com/dovecot/pigeonhole/blob/master/src/lib-sieve/sieve-binary-code.c#L300

Regards,

Stephan.