[developers] "nested unification context"
Emily M. Bender
ebender at uw.edu
Sat May 11 00:27:25 CEST 2013
I suspect the students are all using the TTY interface, as I
usually encourage it. I'll recommend that they leave the error
context when they end up in it.
Thanks,
Emily
On Fri, May 10, 2013 at 3:15 PM, Stephan Oepen <oe at ifi.uio.no> wrote:
> great, thanks, ned!
>
> i can reproduce the issue, though only when parsing from
> the Lisp prompt, i.e.
>
> (lkb::do-parse-tty "der man leyen-t dos bukh")
>
> as someone who programs in lisp, i suspect that is what
> you usually do? if so, the error goes away when i leave
> the error context, i.e. type ‘:reset’ to the lisp prompt.
>
> when going through ‘Parse|Parse input ...’, i cannot get
> nested unification contexts using your grammar; is there
> more to the recipe? or are actually those students who
> have run into this problem all using the TTY interface?
>
> best, oe
>
>
> On Fri, May 10, 2013 at 11:42 PM, Ned Letcher <nletcher at gmail.com> wrote:
>> I've been able to reliably reproduce this -- at least on my machine. This
>> error seems to occur following a parse that was aborted due to running out
>> of edges. ie:
>>
>> "Probable runaway rule: parse/generate aborted (see documentation for
>> *maximum-number-of-edges*)"
>>
>> Once this has occurred it seems that anything that then involves unification
>> results in the nested unification error, such as parsing, or loading a
>> grammar.
>>
>> I've attached a grammar that will give the max edges error if you try to
>> parse "der man leyen-t dos bukh". Then you can reload the grammar or try to
>> parse the string again to get the nested unification error.
>>
>> Ned
>>
>>
>>
>> On Fri, May 10, 2013 at 6:04 AM, Ann Copestake <Ann.Copestake at cl.cam.ac.uk>
>> wrote:
>>>
>>>
>>> in case it helps - I recollect that error message as turning up in highly
>>> misleading contexts. The ones I remember were code bugs - feature
>>> structure
>>> creation not being called correctly. These could potentially turn up in
>>> lkb/user-fns.lsp as well as in the main code, so if that's been modified,
>>> I'd
>>> be suspicious.
>>>
>>> Best,
>>>
>>> Ann
>>>
>>>
>>>
>>>
>>
>>
>>
>> --
>> nedned.net
>
--
Emily M. Bender
Associate Professor
Department of Linguistics
Check out CLMS on facebook! http://www.facebook.com/uwclma
More information about the developers
mailing list