[CWB] cqpserver charset: Where can I set this variable?

Jörg Knappen j.knappen at mx.uni-saarland.de
Fri Apr 4 13:52:49 CEST 2014


I tried the most recent version (3.4.7 Revision 545). It compiled  
smoothly on my
SLES 11 SP 2 (after installing a few *-devel packages, as usual in  
this setting).

I started the cqpserver using the old initialisation file. This went.

But: My client (SRUCQIBridge from CLARIN) was unable to connect to  
cqpserver. It looks
(in the debug output) like the login failed.

I'll attach the debug output at the end of this mail.

--Jörg Knappen


End of things written to STD_ERR with -d ALL. Note that username and  
password are
both 'test' in my setting.

[...]
CQi: Connection established. Looking up client's name.
CQi: ** new CQPserver created, initiating CQi session
CQi: creating attribute hash (size = 16384)
CQi: waiting for command
CQi RECV BYTE 0x11
CQi RECV BYTE 0x01
CQi RECV BYTE 0x00
CQi RECV BYTE 0x04
CQi READ WORD   0004      [= 4]
CQi RECV BYTE[4]
CQi READ CHAR[] 'test'
CQi RECV BYTE 0x00
CQi RECV BYTE 0x04
CQi READ WORD   0004      [= 4]
CQi RECV BYTE[4]
CQi READ CHAR[] 'test'
CQi SEND WORD   0102      [= 258]
CQi FLUSH


Zitat von Stefan Evert <stefanML at collocations.de>:

> On 2 Apr 2014, at 14:52, Jörg Knappen <j.knappen at mx.uni-saarland.de> wrote:
>
>> # corpus properties provide additional information about the corpus:
>> ##:: charset  = "latin2" # character encoding of corpus data
>> ##:: language = "pl"     # insert ISO code for language (de, en, fr, ...)
>>
>> However, the cqpserver still claims (verified using -d ALL) that the corpus
>> in encoded in "latin1". It should announce "latin2" here ...
>
> Yes, the backend function is just a dummy.
>
>> Where does the cqpserver take the character set from, and how can I  
>> modify this?
>
> This has been fixed in the current beta version (3.4 series).  We  
> didn't backport the change, since CWB 3.0 doesn't properly support  
> charsets other then latin1 anyway.
>
> You should consider upgrading to CWB 3.4; you'll probably have to  
> compile it from the SVN repository in order to get a working  
> CQI_CORPUS_CHARSET, since the patch was only added this February.
>
> Best,
> Stefan
> _______________________________________________
> CWB mailing list
> CWB at sslmit.unibo.it
> http://devel.sslmit.unibo.it/mailman/listinfo/cwb





More information about the CWB mailing list