RE: Find/Replace in IFS file - The fastest way to do it -- RPG400-L

Hadn't thought of this!  So are you thinking I should read it into something
that can hold it (i.e. User space) or do some pointer arithmetic on a string
that has a base pointer?  I am thinking it would probably be faster to do
the pointer arithmetic.

I'd use the TS_malloc() API to allocate one piece of memory for the wholething. I'd use TS_malloc() instead of %alloc() because it allows you tocreate an allocatation that exceeds 16mb. I don't know if that's neededor not.

The next problem will be doing the replaces. It's relatively easy tosearch a big area of memory, but much harder to replace the dataefficiently.

The problem is "moving" all of the data that comes after the replacementwhen the original length and the new length aren't the same.

For example, if you have "Press the <enter> key" and you make itsay "Press the <enter> key" how would you do it? You search for <.When you find it, you replace it with "<"... but since "<" is onecharacter and "<" is 4 characters, you'll have 3 extra spaces after the<. How do you get rid of those spaces? You copy the text"enter> key"3 spaces to the left... that eliminates the extra space.


Not a big deal with a string that short, it'll only take a second.

However, if you have megabytes of data, and you have to copy it over a fewspaces with every single find/replace operationg that you do, that'll adda lot of processing time.

You might be able to whip up a nice search and replace function using

memcpy and memcmp.

Also memchr(). You use memchr() to find the initial character in memory.Then you use memcmp() to see if it matches the whole string. Then you usememcpy() (or probbaly memmove()) to do the replace.

In my experiencing, using memchr() followed by memcmp() is much fasterthan using memcmp() on every byte, or checking each byte individuallyusing a based field. YMMV, of course.

Anyway, Aaron, I thought you didn't do RPG anymore? Didn't you say youbecame a full time Java guy?