Re: [HACKERS] grammer/keywords/shift/reduce conflicts - Mailing list pgsql-hackers

From Thomas G. Lockhart
Subject Re: [HACKERS] grammer/keywords/shift/reduce conflicts
Date
Msg-id 34F5015D.30E6F2C6@alumni.caltech.edu
Whole thread Raw
In response to grammer/keywords/shift/reduce conflicts  (Brett McCormick <brett@work.chicken.org>)
List pgsql-hackers
> well, by putting TRANSACTION and ORDER in the ColID grammer, I seem to
> have introduced some shift/reduce and reduce/reduce conflicts.. will
> the grammer work?  What are your thoughts on using these as column
> identifiers?  If they aren't going to end up usable I certainly won't
> use them as table/field names.. (order sounds like a really bad idea)

Yup. I think that the conflicts mean that there now would be ambiguous
grammar. So, if you stumble across just the right statement and order of
words, you may not get what you expected, and not be able to get what you
want. Both "transaction" and "order" are pretty clearly SQL-ish words, so
I wouldn't bother trying to make them work in other contexts...

                                  - Tom


pgsql-hackers by date:

Previous
From: "Thomas G. Lockhart"
Date:
Subject: Re: [HACKERS] mode of libs
Next
From: jwieck@debis.com (Jan Wieck)
Date:
Subject: Re: [HACKERS] SELECT currval('SEQ') broken?