SEGV when exercising code at bottom of processing table

[Barry Wiseman]

Correct, above a certain position near the bottom of the table the errors do not happen.

On 2/5/2010 1:56 PM, John Esak wrote:
> You are thinking it is because it is low in the prc table. I've seen this,
> too.  What would happen if you copy the code to line 1 of the table and
> redirect it to be run from there.  Would the error happen?
> John
>
>> -----Original Message-----
>> From: filepro-list-bounces+john=valar.com at lists.celestial.com
>> [mailto:filepro-list-bounces+john=valar.com at lists.celestial.co
> m] On Behalf Of Barry Wiseman
>> Sent: Friday, February 05, 2010 1:34 PM
>> To: Filepro_List
>> Subject: SEGV when exercising code at bottom of processing table
>>
>> fP 5.0.15 on linux
>>
>> We have an input processing table which is currently 5886
>> lines of code.  (The associated automatic table is 6845
>> lines.)   Several chunks of code near the end of the table
>> are throwing SEGV's when called at runtime.  Moving these
>> code blocks higher in the table constitutes a workaround.
>>
>> The simplest example is:
>>
>> 	@key4:::
>> 	:@sn eq "4":switchto "4t"  ; end:
>> 	::switchto 4     ; end:
>>
>> With this code block located at line 5861, pressing "4" from
>> a record generates a segmentation fault.  When I move this
>> three-line block above the short subroutine which precedes it
>> in the table -- so that it is now located at line 5852 -- the
>> @key4 executes without a problem.
>>
>> Another example, a short subroutine
>>
>> 	ttype:::
>> 	:18 eq ""  or  18 eq "SPOT":return:
>> 	::lookup typ = gottype  k=18   i=A -nx:
>> 	:not typ:return:
>> 	:typ(21) eq "Y":ltagtype = "LTL":
>> 	::return:
>>
>> which works fine when located above line 5856, SEGV's when
>> lower in the table, or, when at the very bottom, generates:
>>
>> 	FATAL ERROR!
>> 	
>> 	Invalid opcode 254 in token table.
>> 	Line 5881, offset 1.
>>
>> Any ideas what the issue could be here?  We have other
>> similar versions of this table which are larger (more lines)
>> but do not produce these errors.
>> _______________________________________________
>> Filepro-list mailing list
>> Filepro-list at lists.celestial.com
>> http://mailman.celestial.com/mailman/listinfo/filepro-list
>>
>
>