Crossfire Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CF: stable/development releases.

On Apr 28, 10:21pm, rhm-crossfire@math.utk.edu wrote:

> Along with this, I have a couple questions.  In the past the release has been
> made in several small pieces.  Is this done for more than download
> convenience?  I think that offering them up like this is perfectly fine along
> with the complete package.  Makes managing your local copy via patches much
> easier.  I would also like to suggest that we start using unified diff for
> our patch type, makes patching easier.

 Can you give a better example of by releases being made in several small
pieces?

 As best as I know, I have been doing the releases pretty much the same way in
the past couple of years (arch, core, map distributions, and more recently
adding doc and client distributions).

 If the question is why doing that instead of one release with everything,
there are a few reasons:

 First, not everyone needs the arch distribution - in fact, the only people
that need the arch are those who wish to fiddle with them or make new ones.  So
that is 900K which is saved on downloading (and unpacking)

 maps:  at 3 megs, they probably don't always need to be included.  In fact,
the need to download every new map release is probably not necessary, as for
the most part, they include bug fixes.

 doc, client:  These are there for those who don't plan setting up a server.

 In practive, the doc, arch, and maps probably don't change enough that there
is any reason to download them everytime.   However, those distributions are
trivially easy for me to make, and I might as well make the latest changes
available.

 the core is more difficult to package, as that requires at least a some
testing (often times not enough, but I try to at least verify that it will
compile on a non linux system).  The doc, arch, and maps are really more than
just a set of files, so there really isn't any testing to do on those.

 I don't know why unified diffs are any easier.  Most all vendor supplied patch
programs will understand context diffs, but not all will handle unified diffs.
 So it could make it more difficult for people to patch the program.  Also, I
am not sure if unified diffs are necessarily any easier to read than standard
context diffs.



-- 

-- Mark Wedel
mark@pyramid.com
[to unsubscribe etc., send mail to crossfire-request@ifi.uio.no]
References: