delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/2002/05/20/18:19:50

From: "deckerben" <deckerben AT freenet DOT de>
Newsgroups: comp.os.msdos.djgpp
References: <3ce7fc1f$0$26970$9b622d9e AT news DOT freenet DOT de> <3CE810BA DOT FCA8CEDF AT phekda DOT freeserve DOT co DOT uk> <3ce8232a$0$26959$9b622d9e AT news DOT freenet DOT de> <3CE93EB2 DOT 2DD59789 AT phekda DOT freeserve DOT co DOT uk>
Subject: Re: ANNOUNCEMENT: Python 2.2.1 for DOS test release
Date: Tue, 21 May 2002 00:13:54 +0200
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Lines: 68
Message-ID: <3ce97174$0$23709$9b622d9e@news.freenet.de>
NNTP-Posting-Host: 213.6.61.201
X-Trace: 1021931893 news.freenet.de 23709 213.6.61.201
X-Complaints-To: abuse AT freenet DOT de
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

> Do you have diffs against the original sources? I can take a look at the
> diffs, if you'd like a second opinion.
>
> (I added a DJGPP port of Python to my list of things to do on Friday,
which is
> a pretty strange coincidence!)

I'm really happy for any advice that is given to me in a positive manor. I
will try to make up a diff package and send it to you. By far the biggest
known bug of the current test release is that running any setup.py script
seems to crash - usually running out of memory when Python tries to start up
GCC.

I have a new version underway, one with the NUMERIC library built in (from
http://www.pfdubois.com/numpy/ ), and some other small fixes. But I want to
wait and see what the score is with the first release before I upload the
next. So I would appreciate feedback. I want to take the package in the
direction of scientific visualization, using various XML implementations as
data-storage format(s). But that's all a long way down the road.

Step-by-step.

> > Actually, I did it because I was looking for a faster alternative to
Batch
> > files.
>
> That's a long way to go, just to get faster scripts! But I can understand
> that.

For 'faster scripts' I realized that I was going to implement a whole new
technology into my system - one that could replace my slow-running batch
files (that need to call a new executable every line) with a
script-interpreter that could do the lion's share under it's own roof. I
decided I wanted to make it count ... so I didn't want to play with some
software toy, but it would need to be a tested, extensible and
industrial-strength solution... preferably one with a lot of support and 3rd
party development.

I considered Perl, Python and Lisp and chose Python.


> I haven't really started learning Python yet. I've done some Zope stuff
and
> I'm going to write some stuff at work in Python, but I'm not at all
familiar
> with the language.

I've been so busy porting the interpreter, I haven't had time to learn it
either :-) But I got some mail from other developers, a few of whom also
started out the same way, by first porting it to their platform.

> The problem with socket libraries is that they only work on a subset of
the
> platforms that DJGPP runs on. That really is pretty inconvenient for the
user.
> Until there's one library for all platforms, I'm not sure it's worth
trying to
> add Socketmodule.

Then, I guess what is needed is a C library (please, NOT TSR) that can
detect a Win95/NT winsock-tcpip connection _if_it_exists. Otherwise, it
tries its own connection via comm port. It needs to have all the Libsocket
definitions, though, so that porting UNIX stuff is made easy. Just my
thinking.

Ben


- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019