RE: Search for SELECT / FROM parameters in a CGI source from RPG

[Jim Langston <jlangston@xxxxxxxxxxx>]

Reading a block at a time may be faster reading, but you get into a REAL
headache with the snippet of the last line you have.

Say we have this data:
Line 1
Line 2
Line 3
Line 4
Line 5
and that each line is terminated with a single end of line char (ASCII has
2, but to make it simpler).  And say we are reading a 15 byte buffer (I
didn't feel like typing to fill a 500 byte buffer <g>).
Your first read would produce:
"Line 1<EOL>Line 2<EOL>L"
and the second read would produce:
"ine 5"
Now you have to come up with some scheme to take the snippet of the last
line, the "L" and add it to the beginning of the block you read.  And since
you only set aside 500 byte for your reads, you'll have to increase your
buffer to 1000 just to be safe, but what if one line is actually 1001 chars
long?  And... It's just a pain in the neck, been there, done that, got the
headache.

Much easier just to read a byte at a time and build your string.  It may be
slower reading, but the code is MUCH easier to maintain and much less prone
to strange bugs.

Regards,

Jim Langston

-----Original Message-----
From: dmosley@dancik.com [mailto:dmosley@dancik.com]

Sounds a bit time consuming.  If you can read one character at a time, can
you specify to pull in a block at a time.  Kind of like, %subst(IFSFILE:
501: 500)

David L. Mosley, Jr.
Systems Analyst
2000 CentreGreen Way
Suite 250
Cary, NC 27513