Re: Does problem NCL-1454 have been solved?

Dear Feng,

[I'm CCing your message to ncl-talk because we're encouraging people to email ncl-talk and not individual developers.]

No, we haven't fixed this problem yet, but I've added your email to the ticket for further reference.

It's at a low priority right now because of the work-around. Let ncl-talk know if this becomes an issue for you.

Thanks,

--Mary

On May 9, 2013, at 10:23 AM, ChenFeng wrote:

> Hi Mary,
>
> I also encounter with the "input in flex scanner failed" problem.
>
> The script works well when I run it manually. However, the error "input in flex scanner failed" comes out when the script runs by the Job management system (qsub). And when I add an "exit" command before the "end", it works well.
>
> I found you have ticketed for this problem from the ncl-talk mail-list. So I want to know whether the problem have been solved?
>
> Regards
>
> Feng Chen
>
>
>
>> Hi Jiapeng,
>>
>> We've seen this issue before, but don't have an easy answer.
>>
>> One user did say he fixed the problem by putting an "exit" command right before the "end" in your main script.
>> We're not sure yet why this fixes the problem in some cases.
>>
>> We have a ticket for this problem, NCL-1454.
>>
>> --Mary
>>
>> On May 16, 2012, at 2:09 AM, 李嘉鹏 wrote:
>>
>>> Hi
>>> I'm using an NCL script to process some data, it takes several hours to complete.
>>> Near the end of the program, NCL aborted with a message "input in flex scanner failed".
>>> There is no clue when or near which line the problem occured, as they seem placed randomly among the program's other output.
>>> I don't know how to reproduce this error message easily since the program would ended successfully last night and this morning with a different set of data files, and I can't afford the time to run the program repeatedly with the current set of data files...
>>>
>>> There's only one place in the source file that contains "input in flex scanner failed", the YY_INPUT macro defined in "ni/src/ncl/Symbol.h".
>>> It looks like a part of program parser which reads the NCL program into the interpreter. (am I right?) So I wonder it might not be some program with the NCL program, but the scanning process. I did modify the NCL program while it was running, at line 169,
>>> g_nearest_model_1d_ind = minind(dist_square)
>>> to
>>> g_nearest_model_1d_ind(si) = minind(dist_square)
>>> Will the NCL interpreter scan the code again after it's done with the data process at line 177 to 191?
>>>
>>> I'm running the program all over again, hoping this time it will pass.
>>> And here's the main program, any suggestion is welcomed.
>>>
>>> Regards
>>> Jiapeng
>>>
>>>
>>> <calc_stat.ncl>_______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu May 9 11:06:37 2013