1. My PEU update
- Posted by Martin Stachon <martin.stachon at worldonline.cz> Nov 21, 2002
- 537 views
To Peu people: I won't be on IRC today, so I am posting it here. We should get out own list. Today I successfully merged most of bits with Menno's latest version (November-2002) from menno.txt: 04 jan 2001 the dir() in windows should use f.write instead of f.create , windows uses unkown for f.create as -1 . in parser automatic change of proctect keywords of C (char,default) in error.c the crash_routine works now also with -bind . I have removed them from builtins.c . the parser had a bind error , a divide should alway set integer_flag to 0 . Other problem with bind is that it sometimes says "not assign a value" it orgin lays in the eval routine but i have not found a solution for it yet ! Also in peuw the binder has no idee what to do with the call_back . 22 sept 2001 Didn't work on peu for a long time , but I speed it up a little bit . namely by autosequence , and the find , equal routine's Somewhere in time I builtin the upper() lower() into builtins.c INSTANCE has been implemented , so we found that there was missing a function F53 . For that reason SCREEN_GET SCREEN_PUT becomes 65 , 66 (change in include\graphics.e) 13 0kt 2001 Change return function result , if it is a slice it didn't destroy it . So it uses a lot of memory . Change the IF statement so it can use the shorthanded form . 14 Nov 2002 Has done a build_sequence for char and words . Safes memory but I have no idea what it does with program speed . I discouverd that it gives a problem by call_proc() call_func() when the parameter sequence becomes a byte sequence it should be convert to a object sequence . So I have comment it out . 17 Nov 2002 Put into it : Martin continue in error.c and changed some of the casting . Put into it : Martin FLUSS (machine_id=60) . ---------------------------- I also fixed Menno's atom_to_float32 and added atom_to_float64, they work ok, but float32_to_atom and float64_to_atom is broken. With Menno's and my changes now most of generic programs that previously crashed now run (mostly because of [s]printf crashing). But Windows is still a problem, \euphoria\demo\win32\winwire.exw crashes and also all win32lib programs ( i think it is c_proc/call_back/... not converting values correctly) Also binding now works better. I cleaned up some non-ANSI code so I hope it compiles ok on non-borland. Also please look at the trick that puts compilation time,date,user into Peu banner (buildinfo.c) The source,binary and some notes are at http://www.webpark.cz/stachon/files/peu.zip Martin
2. Re: My PEU update
- Posted by Euler German <efgerman at myrealbox.com> Nov 21, 2002
- 492 views
On 21 Nov 2002, at 17:42, Martin Stachon wrote: > > To Peu people: > I won't be on IRC today, so I am posting it here. We should get out own > list. > There's an Euphoria list at Coolist already and as far as I know it's on back burner. I think either Rett Williams or jbrown is managing it. As it's already set the only thing to do is subscribing. If anyone wanna look: http://www.coollist.com/group.cgi?l=euphoria -- Euler
3. Re: My PEU update
- Posted by jbrown1050 at hotpop.com Nov 22, 2002
- 547 views
On 0, Martin Stachon <martin.stachon at worldonline.cz> wrote: > > To Peu people: > I won't be on IRC today, so I am posting it here. We should get out > own list. Agreed. Coollist, or Topica? or is there perhaps another alternative? > > Today I successfully merged most of bits with Menno's latest version > (November-2002) Wow! Thats a lot newer than the menno version I got! > from menno.txt: > > 04 jan 2001 > the dir() in windows should use f.write instead of f.create , windows uses > unkown for > f.create as -1 . > in parser automatic change of proctect keywords of C (char,default) > in error.c the crash_routine works now also with -bind . I have removed them > from > builtins.c . > the parser had a bind error , a divide should alway set integer_flag to 0 . > > Other problem with bind is that it sometimes says "not assign a value" it > orgin lays in > the eval routine but i have not found a solution for it yet ! > Also in peuw the binder has no idee what to do with the call_back . > > 22 sept 2001 > Didn't work on peu for a long time , but I speed it up a little bit . > namely by autosequence , and the find , equal routine's > Somewhere in time I builtin the upper() lower() into builtins.c > INSTANCE has been implemented , so we found that there was missing a function > F53 . > For that reason SCREEN_GET SCREEN_PUT becomes 65 , 66 (change in > include\graphics.e) > > 13 0kt 2001 > Change return function result , if it is a slice it didn't destroy it . > So it uses a lot of memory . > Change the IF statement so it can use the shorthanded form . > > 14 Nov 2002 > Has done a build_sequence for char and words . > Safes memory but I have no idea what it does with program speed . > I discouverd that it gives a problem by call_proc() call_func() when the > parameter > sequence becomes a byte sequence it should be convert to a object sequence . > So I have comment it out . > > 17 Nov 2002 > Put into it : Martin continue in error.c and changed some of the casting . > Put into it : Martin FLUSS (machine_id=60) . > > ---------------------------- > > I also fixed Menno's atom_to_float32 and added atom_to_float64, they work > ok, but float32_to_atom and float64_to_atom is broken. I implemented my own version of atom_to_float32 and float32_to_atom (thanks to mic_!) but it hasnt been tested yet. > > With Menno's and my changes now most of generic programs that previously > crashed > now run (mostly because of [s]printf crashing). But Windows is still a > problem, > \euphoria\demo\win32\winwire.exw crashes and also all win32lib programs ( i > think it is > c_proc/call_back/... not converting values correctly) > > Also binding now works better. > > I cleaned up some non-ANSI code so I hope it compiles ok on non-borland. > > Also please look at the trick that puts compilation time,date,user into Peu > banner > (buildinfo.c) > > The source,binary and some notes are at > http://www.webpark.cz/stachon/files/peu.zip > > Martin > I'll look at the source, and perhaps try to merge them ... it would be easier if we had CVS though .. *sigh* jbrown