Re: Release philosophy (was: Re: Miscellaneous Things)

[Eric McDonald <mcdonald at phy dot cmich dot edu>]

Hi Hans et al.,

On Sat, 6 Dec 2003, Hans Ronne wrote:

> has a good point, however, in that more snapshots/release candidates are
> needed in the final phase before a release. I think Eric's Windows
> Installers and rpm packages can do this job, but they could perhaps be more
> frequent.

I agree about the frequency. I have been waffling back and forth 
about whether to wait for the rest of the path-related changes, 
but since that might be awhile, I am hoping to release tonight or 
else tomorrow. Since the last release, we have at least two major 
improvements: your fixes of the module loading code and a new 
version of the pathing code (with desync fix :-). I am hoping to 
get Unix and Windows file naming over to the new nomenclature 
which we agreed upon. I have this mostly done; just need to test 
and to figure out how I want to handle old prefs files. Also, I 
guess we have a bunch of new game modules that people might want 
to play (thanks Lincoln and Elijah)....

I probably should have released after you took care of that module 
loading stuff....

Wrt to RPM's: I think I am going to split xconq.spec into 
xconq.spec and xconq_cvs.spec (or xconq-cvs.spec, whatever the 
generally accepted RPM naming convention is as far as CVS 
snapshots are concerned; I have not researched this yet). The 
xconq_cvs RPM's would be versioned by date, and the release RPM's 
would be versioned by, __well, by version.

Also, I think I am going to release just a plain tarball and/or 
zip file of the Xconq files minus any binaries. This may be useful 
to some people. Probably means that I should take a look at the 
'dist' target in the top-level makefile....

  So much to do, so little time,
  Regards,
    Eric