[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Axiom-developer] Release?

From: root
Subject: Re: [Axiom-developer] Release?
Date: Sat, 11 Dec 2004 13:57:54 -0500


I haven't tried the new axiom exe files yet (I have to secure access
to a windows box and I just found out about it at midnight last night).
Assuming it works (and even if it fails we fix it) there is no possible
objection to announcing it. It appears that the .exe file does not do
a system build so it doesn't require any development tools.

It is clear there is a widespread interest in the windows version just
from the number of downloads so far. So it is a good idea to try and
meet this demand as soon as possible. 

I'm not sure what "announcing" it really means. Clearly I've made no
effort to build up a "marketing and sales" channel for axiom so there
is no obvious place to "announce". You're welcome to take the lead in
this (aka "director of marketing" :-) ) and figure out a way to market
axiom. I'll provide what help I can but I'm a complete novice at this.

Perhaps the easiest way would be to put up wiki pages. I'm willing to
be the test case for these pages as I'm going to play with the .exe
file shortly.

IBM, RedHat, and Red Flag Linux (China) have all announced "ISV
certification programs" that are intended to certify that programs
work on their respective platforms. Perhaps an effort to contact
these people would get us much more exposure much faster as well as
resources to build, test, and support axiom. The Red Flag Linux 
program includes "propaganda" as part of the package.

Merging the code base into the main branch is significantly more work.
Several questions come to mind...

Will the axiom--windows--1 branch build cleanly on windows?

  by this I mean that I can download the tla file to a freshly
  installed windows box and build it.

  what is the procedure? it should work by typing 'make'

  is there a tla for windows?

  what software must the user install on the box? 
  clearly some form of gcc and latex plus awk, grep, rm, etc.
  what is MSYS and where can I get it? should we package it in zips?

  what does Windows use to read dvi files? should we cross-compile
  and package advi (which is already in zips?) or should the windows
  build automatically create pdf files?

  is the version of gcl on the axiom--windows--1 branch the one that
  should be used with windows? does it also build with linux?

  is there a <<Makefile-windows>> stanza in the top level Makefile.pamphlet?

  I haven't been involved in any of this so I'm unclear about what
  exact steps are required.

  I don't yet have the ability to support questions about this branch
  as I don't yet have access to a windows machine.

  there appears to have been an issue about file names and path names
  have these issues been resolved?

beyond the algebra portion comes the question of sockets

  will it compile cleanly under the windows MSYS environment?
  does the axiom socket code work on windows?
  does sman compile?

and the question of X compatibility

  does the graphics code compile?
  does it run?

on linux the graphics exists, sman exists, and hypertex is nearing
completion (it segfaults but I'll find it eventually). have we looked
at getting these branches to compile?

I did spend two hours in the bookstore yesterday reading thru a book
on windows sockets. They look very similar but I'm certain there are
compiling issues. gcc won't know about windows sockets.

I'll diff the two branches (main and windows) and merge as much as
I can test. And if you'll set up an instructions page on Page's site
or send me instructions I'll set up a test environment on the windows

I'm rather reluctant to merge the code bases until I can assure 
myself that I can build the system on both platforms and test it 
from the sources. It's not apparent to anyone but I'm nearly 
always running a system build and tests on some machine or other.
Merging the code and pushing it out to savannah isn't an issue
for people who want to just use it. The savannah code needs to 
be as clean as we can make it so new developers aren't struggling
with poor code. It's a quality issue.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]