Greetings ooRexx'ers,
Gilbert Barmwater wrote:
And, yes, I know that stems are NOT ordered collections!
Well Stems may operate in two modes...
1) Ordered where .0 is the number of stored items and increment .# is the
ordered storage locations
2) Sky's the limit!
And Rony, I remind you that there are those old methods probably dating back to
Object Rexx v1.x that had very poor performance. Glanging at my Text File class
changelog....
/* FILENAME : LDSTextFile.cls
*/
/* TYPE : Open Object Rexx - OS/2, Win32, Linux, AIX
*/
/* DESCRIPTION : Class for handling text files
*/
/*
*/
/* AUTHOR : Michael D Lueck
*/
/* mlu...@lueckdatasystems.com
*/
/*
*/
/* NEEDS : LDSFileSystem.cls
*/
/* LDSRexxUtil.cls
*/
/*
*/
/* USEAGE :
*/
/*
*/
/* REVISION HISTORY
*/
/*
*/
/* DATE REVISED BY DESCRIPTION OF CHANGE
*/
/* ---------- ----------
------------------------------------------------------- */
/* 1999/10/31 MDL Initial Creation
*/
/* 2000/06/04 MDL Change internal FINDarray from an array to a stem to
*/
/* support IBM extended SysUtil stem functions
*/
/* 2000/08/08 MDL Change to add GETlogging to set the debug/verbose
flags */
/* 2000/10/25 MDL Moved STEM API's business logic to LDSRexxUtil
*/
/* 2000/10/27 MDL Added Test method for proper init error detection
*/
/* 2000/12/18 MDL Added Empty Method
*/
/* 2001/01/06 MDL Changed empty lines to =0
*/
/* 2001/06/17 MDL Changed to use Array objects vs Stem objects for
memory */
/* 2002/07/11 MDL Fixed bug in INSERTline where it was still calling a
stem */
/* related method vs the array related one
*/
/* 2004/05/11 MDL Change to READfile to use new HighPerf I/O option
*/
/* 2004/05/12 MDL Added support for global trace options within Object
CUR */
/* 2004/09/09 MDL Change to WRITEfile to use new HighPerf I/O option
*/
In 2001 I switched from Stem based storage to Array based storage. So I suspect
there was old ORexx v1.x methods to read and write from files to Stem objects
which performed poorly, so I had to
abandon Stem storage to seek after better performance.
Then IBM added some High Performance I/O methods for accessing Array objects....
From READfile:
/* Check for how many lines the file has */
FILEsize = file~command('QUERY SIZE')
/* Read in the file */
/* This code requires at least ORexx 2.1.1 on Win32 - only tested with
>=2.1.2 */
INstr = file~charin(1, FILEsize)
FINDarray = INstr~makearray()
From WRITEfile:
/* Write out the file */
/* This code requires at least ORexx 2.1.3 on Win32 */
if file~LineOut(FINDarray~MakeString()) then do
LDSLogging~Logger('OBJECTCUR LDSTextFile~WRITEfile -> Error calling
LineOut, RC=' || rc)
file~close()
return 0
end
>>> So seems the added methods were the MakeArray() method of the String class and the
MakeString() method of the Array class. <<<
Man.... where did time go.... 1999 to 2004.... 20+ years have transpired!
Anyway, evidently there were some ORexx API's meant to map in/out of Stem
objects, which were the poor performers. Since the topic is making API's to
bridge between files and ooRexx collection
classes, I thought to revisit what was there and not performing well, back
then. Hopefully that old ORexx v1.x code might be improved. I would not mind in
the least being able to interchangeably use
an Array or a Stem storage class. That would be cool! (+1) I just do not think
the ooRexx interpreter should impose a significant performance penalty for
using an Array versus a Stem to store lines of
a Text file.
I am thankful,
--
Michael Lueck
Lueck Data Systems
http://www.lueckdatasystems.com/
_______________________________________________
Oorexx-devel mailing list
Oorexx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oorexx-devel
