Re: PQunescapeBytea code

Adam Kavan wrote:
> Actually I was looking at that code today and it does not ignore something 
> if it is escaped by a backslash on not on the list.  It eats the backslash 
> and then continues the loop so next time that character will be parsed 
> normally.  However PQunescapeBytea is _very_ slow.  I am storing fairly 
> large (several hundered K) byte strings into Bytea's and it can take 30 
> seconds or more to convert them back into binary data.  I wrote a new 
> version of PQunescapeBytea that uses pointers instead of arrays to store 
> the string in, this increases the speed about 30 fold on my strings and 
> still has the same behavior.  I wasn't sure if this would be something I 
> should submit as a patch or not, is anyone interested in this?
> 
> If they are I'll try to figure out how to submit a patch.

Are you testing againts 7.3.X or 7.4?  7.4 has a faster version.  If you
are testing against 7.4, do a diff -c against the old and new files and
send it to the patches list.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
359-1001+  If your life is a hard drive,     |  13 Roberts Road +  Christ can be your backup.        |  Newtown Square,
Pennsylvania19073