Lutz:
Since I write newlisp programs on Linux and run them on windows. The Nan and Inf discussion and differences between Win32 and Linux versions caught my attention.
Refresh my memory. Why did you start a separate Windoze version of NewLisp again?
By the way, the GUI stuff in the old Win versions was the about the easiest GUI stuff I have ever worked with. Although I understand why you did away with it, I hated to see it go.
Eddie
Win version
I agree! I was looking for a language that had extensions for networking and GUI programming on the Windows platform, that was NOT C++, or Java. I guess it's a personal thing, but I just can't think in that family of languages. Forth would have been perfect, but there are no implementations of Forth that will adequately do both.
Lisp comes very close to Forth in its elegant simplicity. Lutz' newLISP blew me away in how easily the Windows GUI and other OS functions could be tamed. I too was unhappy that Tk was used to replace the excellent newLISP hooks into the native functions, though I agree it is probably a better solution overall.
Lisp comes very close to Forth in its elegant simplicity. Lutz' newLISP blew me away in how easily the Windows GUI and other OS functions could be tamed. I too was unhappy that Tk was used to replace the excellent newLISP hooks into the native functions, though I agree it is probably a better solution overall.
why Windoze version again
Some users wanted it, some useres had problems running Cygwin based newLISP and other Cygwin stuff.
The final push came from Steve who delivered a port to the Digital Mars Compiler and than later to the Borland Compiler. Steve put a lot of good work and research into it, and I think it was worth it. As a former Borland employee I also have some emotional attachements to their tools and their quality ;-)
Using the Borland compiler the changes where relatively small (in the beginning), but over time more and more stuff is creeping in to mask differ3ences between the Cygwin/Unix versions.
But still I think for those who want run Win32 and are afraid of Cygwin hassles the Win32 version has some advantages over the Cygwin version, like only one file, same commandline behavior as the 'Dos' box etc..
Although over time I am trying to minimize differences between the versions, I still recommend the Cygwin version on Win32, if you must work multiplatfrom, i.e. write newLISP code running on various OS's
On the UNIX/Cygwin versions:
(1) sockets can be used as files handles
(2) process stuff with exec (former read/write-process) work fine
(3) data/time stuff is identical
(4) no RET/LF versus LF confusion (drives me nuts when doing web stuff :( )
See the new development version: http://newlisp.org/download/development/
for some further fixes for the Win32 version
Lutz
The final push came from Steve who delivered a port to the Digital Mars Compiler and than later to the Borland Compiler. Steve put a lot of good work and research into it, and I think it was worth it. As a former Borland employee I also have some emotional attachements to their tools and their quality ;-)
Using the Borland compiler the changes where relatively small (in the beginning), but over time more and more stuff is creeping in to mask differ3ences between the Cygwin/Unix versions.
But still I think for those who want run Win32 and are afraid of Cygwin hassles the Win32 version has some advantages over the Cygwin version, like only one file, same commandline behavior as the 'Dos' box etc..
Although over time I am trying to minimize differences between the versions, I still recommend the Cygwin version on Win32, if you must work multiplatfrom, i.e. write newLISP code running on various OS's
On the UNIX/Cygwin versions:
(1) sockets can be used as files handles
(2) process stuff with exec (former read/write-process) work fine
(3) data/time stuff is identical
(4) no RET/LF versus LF confusion (drives me nuts when doing web stuff :( )
See the new development version: http://newlisp.org/download/development/
for some further fixes for the Win32 version
Lutz