Skip site navigation (1) Skip section navigation (2)

Re: StringInfo Macros

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Markus Wanner" <markus(at)bluegap(dot)ch>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: StringInfo Macros
Date: 2009-08-28 14:35:21
Message-ID: 8511.1251470121@sss.pgh.pa.us (view raw or flat)
Thread:
Lists: pgsql-hackers
"Markus Wanner" <markus(at)bluegap(dot)ch> writes:
> trying to clean up the Postgres-R code further, I would like to make  
> use of StringInfo and accompanying functions in libpq/pqformat.c  
> instead of some home-brown duplicate of it. However, I'm missing  
> helper macros like these (mainly for readability of the code):

> #define BUFFER_START_PTR(buffer) ((buffer)->data))
> #define BUFFER_END_PTR(buffer) ((buffer)->data + (buffer)->len)
> #define BUFFER_CURSOR_PTR(buffer) ((buffer)->data + (buffer)->cursor)

> #define BUFFER_BYTES_FREE(buffer) ((buffer)->maxlen - (buffer)->len)
> #define BUFFER_BYTES_READABLE(buffer) ((buffer)->len - (buffer)->cursor)

> Is there any compelling reason these don't exist (and shouldn't get  
> added)? IF not, I'd be happy to create a patch to add and make use of  
> such macros.

Don't call them BUFFER_FOO --- that term already has a pretty specific
meaning in most of the backend.  STRINGINFO_FOO is a bit long but
seems to fit the best with the existing naming in stringinfo.h.

A number of the specific things you are proposing above seem to be
violations of the abstraction stringinfo is meant to present.
It's a string, and the fact that there's any extra space beyond
the end of the string is an internal matter.  In particular I wonder
exactly what you think END_PTR or BYTES_FREE would be used for.

Also, the cursor field is meant to be manipulated directly by calling
code, so I'm not sure how much value there's going to be in abstracting
that.  CURSOR_PTR as shown above seems a bit dangerous --- it might
encourage someone to hold onto such a pointer across a StringInfo
operation that could move buffer->data.  Although maybe I'm overthinking
that, since someone could try that with or without a macro :-(

			regards, tom lane

In response to

Responses

pgsql-hackers by date

Next:From: Werner EchezuriaDate: 2009-08-28 14:39:49
Subject: Re: return a set of records
Previous:From: Kevin GrittnerDate: 2009-08-28 14:30:51
Subject: Re: patch: Review handling of MOVE and FETCH (ToDo)

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group