[CWB] Assertion failed

[Ruprecht von Waldenfels]

Hi, yes, with the same effect. It does access the file, as I said. If I 
replace the corpus name with one the does not exist, I get an earlier 
error message:
cl_new_corpus: can't locate <bulgakovmaster_rsu> in 
/data/parallel/ParaSolFeb14/Registry
Can't access corpus BULGAKOVMASTER_RSU (aborted) at 
/usr/local/bin/cwb-align-export line 66.

This is on line 66, while the error message I get with a valid corpus 
name is further below.
Best, Ruprecht

Am 04.01.2016 um 21:27 schrieb Hardie, Andrew:
> Looks like CQP is reading the wrong registry directory: have you tried it with an explicit -r ?
>
> Andrew.
>
> -----Original Message-----
> From: cwb-bounces at sslmit.unibo.it [mailto:cwb-bounces at sslmit.unibo.it] On Behalf Of Ruprecht von Waldenfels
> Sent: 05 January 2016 01:56
> To: cwb at sslmit.unibo.it
> Subject: Re: [CWB] Assertion failed
>
> Dear Stefan,
>
> unfortunately, I still get an error:
> /opt$ cwb-align-export -o out.txt BULGAKOVMASTER_PL BULGAKOVMASTER_RU s
> {s_id}
>
> =+===CWB::CQP ERROR=====
>    | CQP Error:
>    |     Corpus ``BULGAKOVMASTER_PL'' is undefined
> =+== occurred at /usr/local/bin/cwb-align-export line 140.
>
> The files do exist, but somehow CQP cannot access them. If I change the
> corpus name to something that REALLY does not exist, cwb-alig-export
> throws an earlier error.
>
> Could this be somehow linked to different CWB versions? All corpora are
> encoded with the 3.4 versions. I don't know what the perl modules refer
> to .
>
> Best, happy new year, and thanks!
>
> Ruprecht
>
>
> Am 18.12.2015 um 05:20 schrieb Stefan Evert:
>>> On 18 Dec 2015, at 13:06, Ruprecht von Waldenfels <ruprecht.waldenfels at gmx.net> wrote:
>>>
>>> I have to say this is after I copied the alignments from a different version of BULGAKOVMASTER_RU. Maybe this is somehow connected. I am not sure it all worked as it should, there is a glitch with the original alignments with this corpus that I still have to look into in some detail.
>> That would easily explain the problem.  If the file you copied references a different version of the corpus, the last alignment region might well extend beyond the end of the new corpus – which would trigger this assertion.
>>
>> Of course, most of the other alignment regions will be wrong, too, but CQP can't know that.
>>
>> The only safe way of copying alignments is to cwb-align-export them into beads using ID references to sentences, and then cwb-align-import in the new corpus.  Or, if you cwb-align-imported them in the first place, simply go through the same steps for the new corpus.
>>
>> I'm very happy to announce that CWB4/Ziggurat will not allow you to copy index files to a different corpus under any circumstances. :-)
>>
>> Best,
>> Stefan
>> _______________________________________________
>> CWB mailing list
>> CWB at sslmit.unibo.it
>> http://devel.sslmit.unibo.it/mailman/listinfo/cwb
> _______________________________________________
> CWB mailing list
> CWB at sslmit.unibo.it
> http://devel.sslmit.unibo.it/mailman/listinfo/cwb
> _______________________________________________
> CWB mailing list
> CWB at sslmit.unibo.it
> http://devel.sslmit.unibo.it/mailman/listinfo/cwb