[Axiom-developer] RE: windows/linux coherency

[Page, Bill]

On Sunday, December 19, 2004 10:41 AM address@hidden
wrote:
> 
> I would like to complete the new version of Rosetta CD for 
> Windows by adding Axiom + WinTeXmacs, executable directly from
> the CD. In order to do it in the best conditions, I think that
> it would be desirable, as suggested by Tim, to provide 
> windows/linux coherency. For this I would propose the following:
> 
> 1 - READ-ONLY DIRECTORIES (on cd-rom):
> 
> Programs\alg\Axiom\v3.0
>                      \axiom.ini
>                      \bin\axiomsys.exe
>                      \(other files of axiom binary distrib)
> Programs\edit\TeXmacs\v1.044
>                      \plugins\Axiom\bin\tm_axiom.exe
>                      \(other files of WinTeXmacs binary distrib)
>

I like the idea of the axiom.ini file. I think as someone
else mentioned earlier, this might be a good way to get
around some of the awkwardness of the setting of the
AXIOM environment variable.

BTW, have you confirmed that it is possible to execute these
programs directly from CD? What about setting the AXIOM
environment variable and the PATH which is done now during
installation? I think TeXmacs might require a place to write
temporary files. I don't think Axiom requires this.
 
> 2 - INITIALIZATION FILES:
> 
> A - Defaults :
> 
> a - axiom.ini in Axiom main directory on cd-rom.
> b - If axiom.ini is present in user's directory, it
>     overrides default axiom.ini. More precisely:
> - On Linux, if $HOME\axiom.ini is present, it overrides 
>   default axiom.ini
> - On Windows, if  USERPROFILE\axiom.ini is present, it 
>   overrides default axiom.ini; if not present, if
>   ALLUSERSPROFILE\axiom.ini is present, it overrides 
>   default axiom.ini
> 
> B - Customized:
> If axiomsys is invoked with the -i switch (or --init, or 
> whatever switch best suited) :
> 
> axiomsys -i My_personal_directory\my_personal_initialization_file.ini
> 
> where the dir name and file name are any legal name for
> the OS, then the file my_presonal_initialization_file.ini
> overrides all other init files.
>

Specifying an alternate axiom.ini file on the command line
as you suggest above would require some serious surgery to
the way Axiom starts up. Axiom already has the ability to
read a file containing Axiom commands on startup called

  axiom.input

I know this works under Linux when the file is placed in
the user's home directory. Under Windows exactly where to
put this file is not too clear. I just tried with a file
of this name in the current directory when starting AXIOMsys
from the command line, but the commands in axiom.input
where not read by AXIOMsys :( So this needs more testing.

Typically this file contains commands like

  )set message autoload off

etc. But perhaps this could be extended to include user
interface options without disturbing Axiom itself. E.g.

  --Option: TeXmacs line-break off

would be treated as a comment by Axiom but could be
read and interpreted by the tm_axiom program on startup.
 
> Right now, axiom.ini would contain :
> - anything that would seem important for Axiom initialization,
> - a switch allowing / preventing Bill's tm_axiom to use 
>   linebreaker.c. This would be desirable to provide flexibility
>   to the user, in case it would happen that either TeXmacs or
>   texbreaker.c can be broken from time to time. 
>   Right now, the switch would be to disallow the use of
>   texbreaker
> 
> 3 - TM_AXIOM
> 
> Introduce in Bill's tm_axiom.c the possibility to select the 
> correct ini file for Axiom as defined above, and to read the
> value of the texbreaker switch.
> 

Yes, this would be quite easy. But since tm_axiom is a separate
program form AXIOMsys and in fact is invoked by TeXmacs, not
by Axiom, the Axiom command line option would not be of much
use here. Perhaps there a similar option already available in
TeXmacs? In that case some change would have to be made to
the scheme coding that actually starts tm_axiom in order to
pass the file name and/or options to tm_axiom. I suppose this
would not be hard for someone who knows scheme and also
something about TeXmacs internals. Is anyone willing to help
me to implement this?

Regards,
Bill Page.