From: Jonathan B. <jbr...@ea...> - 2004-08-15 19:59:42
|
On Sun, 2004-08-15 at 14:50, Martin Costabel wrote: > > Perhaps Mr. Costabel would be willing to add it in the fink package's > > post-install script? (see /sw/lib/python2.3/idlelib/config-main.def) > > I have started to look at this, and I have a question: Right now, the > idle_VPython module is not being installed (contrary to what the new web > page seems to indicate). This is the default for python-2.3 as it seems, > because there is already an IDLE module. Would it be useful to install > this module or is it not advisable with python-2.3? Is it too broken? Way back in the day, VPython's original author forked idle to make idle_VPython. Sometime between then and now, the "idlefork" project was created based on Mr. Sherer's work. After enhancing IDLE further, idlefork became the stock IDLE in Python 2.3. Since IDLE in Python 2.3 is essentially idle_VPython + tweaks, we don't install idle_VPython on Python 2.3+. I'll bet that idle_VPython suffers from the win vs. darwin bug as well, but I haven't looked. > I don't know much about IDLE, and there is still a bug in Fink's python > package that prevents IDLE to read html documentation (it thinks > dar'win' is a kind of windows OS). This is not hard to fix, but I am not > the maintainer of the python package. Please point fink's IDLE maintainer at this: http://sourceforge.net/tracker/index.php?func=detail&aid=900580&group_id=5470&atid=105470 It is a bug that I filed against IDLE upstream that lead to changes in Python CVS. I haven't tried it yet (actually, I just found out that it was closed). > The advantage of installing > idle_VPython would be that I could easily configure this so that it > finds the documentation and maybe fix the bug if it is there, too, or > other bugs. Well, since idle_VPython isn't really maintained anymore, I wouldn't recommend packaging it separately, but that is your choice. If you do decide to install idle_VPython anyway, see the options --{en,dis}able-idle-vpython and --with-idle-vpython-dir= in configure. > I have also been looking at vpython-3.0 on and off, too. It is not too > hard to translate the mkdist_osx.sh script into a Fink package > description. What keeps me from releasing a Fink package is that it > needs a complete gcc and boost, and I am not yet clear in my mind how > this should be related to other Fink packages. Is it acceptable that > building vpython would use 600 MB of disk space and take 5 hours? I was > thinking of splitting it and have a separate gcc package, but this > needs some more thinking. I understand. Since 3.0's changes are mostly internal, there's no rush to upgrade yet. Something to think about for the future: VPython 4.0 will use Gtkmm 2.2 + Gtkglextmm 1.0 on UNIX-like platforms. I note that neither of these are in fink right now, but both of them build out-of-the box on OS X when the Gtk C development libs are installed from fink. Although 4.0 will include new user-visible features, even a public alpha release is some ways off, so no rush there, either. -Jonathan |