RE: antelope errors

["Al Honey" <ahoney@keck.hawaii.edu>]

Well, I’m not going to relax yet because the first example I gave was tiny but this one is not:

 

sh ../makeVersion ../Version > sncVersion.c^M

rm -f snc.c y.tab.c^M

rm -f snc.h y.tab.h^M /usr/local/epics/r3.13.10/base/bin/solaris/antelope -d -v ../snc.y^M

/usr/local/epics/r3.13.10/base/bin/solaris/antelope: 439 shift/reduce conflicts,

189 reduce/reduce conflicts.^M

mv y.tab.c snc.c^M

mv y.tab.h snc.h^M

 

There are so many conflicts and they are pertinent to the sequencers!

 

AH

 

 

-----Original Message-----
From: Ralph Lange [mailto:Ralph.Lange@bessy.de]
Sent: Wednesday, March 22, 2006 1:21 PM
To: Al Honey
Cc: tech-talk@aps.anl.gov
Subject: Re: antelope errors

 

Hi Al,

relax.

These conflicts occur when the grammar that antelope builds the parser for has ambiguities. At the point where antelope realizes there might be a situation where the parser has two choices in interpreting the input, it issues a warning and makes the parser deliberately select one of the choices.
The IBM Aix documentation has a nice one page explanation for what's happening.

Of course one could try to change the grammar to remove the ambiguities, but the conflicts occur most often in places where you really don't care and both choices would be working anyway, so ...

Hope this helps,
Ralph


Al Honey wrote:

Aloha

 

This is more ignorance on my part. Do I need to track down the cause of the warnings in the input file or do I ignore these as there is a function call to remove_conflicts() (i.e. does that function fix things or delete the stuff that was found to be errant from the output?)? If the input file needs to be fixed is there a mechanism for enabling more error message so as to find the lines in question?

 

/usr/local/epics/r3.13.10/base/bin/solaris/antelope -t ../pvload.y

/usr/local/epics/r3.13.10/base/bin/solaris/antelope: 4 shift/reduce conflicts, 5

reduce/reduce conflicts.

mv y.tab.c pvload.c

 

Thanks for any help

 

AH