Greetings Rony, Rony G. Flatscher wrote:
What I understood was that you had an ooRexx program that performed poorely in the past and that you were not sure whether that problem is still present. Therefore I tried to ask you to test that program against modern versions of ooRexx to see whether your observation of more than ten, fifteen years ago would still hold.
No, back in 2001 I had to switch from Stem to Array storage due to very poor performance on ORexx accessing Stem objects. Back then the performance of adding to an ORexx collection object grew worse at an exponential rate as each new entry added to the collection required ORexx to re-dimension the storage container. That was a large pain point to migrate code from Classic Rexx on OS/2 to Object Rexx on Windows!
If you do not experience any performance problems with interacting with streams/files, then I would regard this thread to be closeable. This is up to you, of course.
I think I will benchmark my Array performance that IBM enhanced back in 2004 in the high performance alternatives they provided as compared to the Stream to Array code.... see if the Stream to Array still perform poorly. As far as the subject line of this email thread.... "RFC: suggesting new methods for .Stem" then still a high performance way to read/write text files to Stem objects would be greatly appreciated! Maybe a corresponding MakeStem method on the String class: FINDarray = INstr~MakeArray() And MakeString method on the Stem class: file~LineOut(FINDarray~MakeString()) 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
