Re: [Texmacs-dev] Win32 Experimental Version

[david]

Nice to see that things are progressing on the win32 front. By
attacking the enemy on multiple fronts we are increasing our chances
at successful world domination :-)

On Wed, Apr 02, 2003 at 01:55:34PM -0400, Dan Martens wrote:
> It should most likely work on Windows 9X, however, in my experience,
> there are minor API changes in the Win32 layer between 9X and NT.
> This may cause exceptions or seg faults when running. Most calls
> used in TeXmacs however are graphics primitives, which are pretty
> much standard between versions. Windows 9X are horrific operating
> systems and are very unstable at best. Most people have switched to
> XP by now, which is quite a good operating system.

FYI, winXP and winNT are mostly IBM's OS/2 with added MS cruft. That
explains that MS engineers seem to have suddently become compentent at
OS design :-P

> I did write some code for profile handling on 9X systems, so that
> should port over easily as well. If anyone has run this on 9X,
> please let me know.

I am going to get a PC with a win2k license (I plan to run it on top
of VMware). Is win2k a brand of win9x or of winNT?


> CVS is not really necessary as I have my own source control
> implemented here. I would prefer to be the only one with access to
> the X layer source right now as well until I get more bugs sorted
> out. It would make it far too confusing if other developers were
> writing code in the X layer.

I would like to know if there is any free (as in GNU) "intelligent"
version control system around?

I had a recent discussion which convinced me that CVS (or Subversion)
is too rigid to handle the development of the TeXmacs mainstream. But
RMS will probably bust our head off if we start using BitKeeper (and
personnaly, I would not like it very much).


> Yes, MikeTeX is huge. It can be seperated quite easily from the
> windows version. All that was changed to this was some of the
> calling conventions to makepk and maketfm.

Maybe you got some insight which may explain the problem ppl are
experiencing with the font generation of TeXmacs?

One some systems, the fonts are not found, and on other systems the
.pk are generated in the current directory. I suspect that TeXmacs
does not interface very cleanly with METAFONT.


> Please do, and let me know what you think. It may not run very well
> on slower computers, as the graphics are a little choppy. This is
> due to 2 factors. 1) The X Emulation layer has to mimic the X event
> system. Due to some inherent problems with Windows messaging
> procedures, this can slow down the program at times of heavy
> messages.
> 2) All texmacs widgets are natively drawn, and do not use standard
> windows widgets. In the future, a REAL windows version may be
> desired, or a port to GTK.

I hope we will be able to do so.

The plan would be to use some free library as a wrapper around the MFC
mess, because no-one enjoys maintaining MFC-bound code.

We have considered various solutions and it did not seem possible to
find a library which is portable (not Qt, because of licensing),
flexible enough (not wxWindows according to Joris) and provide native
look and feel (not GTK).

                                                            -- DDAA