Re: package proposal update: suite3270

[Corinna Vinschen <vinschen at redhat dot com>]

On Fri, Oct 31, 2003 at 05:17:29PM -0800, Peter A. Castro wrote:
> On Fri, 31 Oct 2003, Corinna Vinschen wrote:
> > > http://www.fruitbat.org/Cygwin/suite3270/setup.hint.suite3270
> >
> > The requires line uses the wrong package names.  The leading "suite3270-"
> > is missing.  Besides that, if the base package already requires to
> > download all other packages, the split would be unneeded.  Did you
> > actually intend that?
> 
> Ok, I knew there would be some confusion concerning this.  Here's the
> rational: The package "suite3270" is actually a super-package (Hmmm...
> [...]
> I'd changed the package names to prefix with 'suite3270-' to group them
> togther, to make them easier to find, but perhaps that wasn't a good idea
> after all.  You can install each product independent of each other.

But that's not the problem.

> So, why isn't the source for each product with each package, you ask?

I never asked this.  This is a good thing, IMHO.

> So, why the "-common" package, you ask?  As I said, each product is
> independent of each other.  However, they all (except pr3287) install
> some common files, which, if you installed the individually would overlay
> each other, and upon removal of one package would remove those common
> files for all.  So, the build process moves those files which are
> identical (common) into a separate package which is prereq'ed by each
> package.  It seemed like a good idea at the time.

It's still a good idea.  I just don't think the super-package idea is
a good one.  What do we have?  5 packages with a common set of files.
So why not keep the common files together in the base package, which
then is not a *super* package but just the container for the common
files and the name and container of the source package.

> > Another nit concerning the documentation.  Each package creates its own
> > documentation subdirectory right below /usr/doc.  So after installing
> > all packages, you have
> >
> >   /usr/doc/suite3270-3.2.20          --  empty!
> >   /usr/doc/suite3270-common-3.2.20   --  empty!
> >   /usr/doc/c3270-3.2.20/
> >   /usr/doc/pr3287-3.2.20/
> >   etc.
> 
> (*sigh*) ... and correct the directory names, but this is starting to get
> out of hand.  I'm starting to think that naming the packages with a
> prefix was a bad idea.  Perhaps having them grouped together by name
> isn't all that helpful.

I don't think so.  Grouping them together using the suite3270 prefix
is a good idea, IMHO.

> > I would prefer to keep all documentation in one subdirectory
> >
> >   /usr/share/doc/suite3270-...
> > and all the above subdirectories below that, instead of polluting the
> > doc directory itself with so many subdirs for one base package.
> >
> > I would also prefer to have only one common README file under
> > /usr/share/doc/Cygwin.  Basically all these READMEs are the same, with
> > just tiny differences.  Why not just one file which describes the
> > whole suite?
> 
> Hmmm... Well, since it's a requirement for each package to have a README,

This is a rule of thumb, not a slavish one.  The package is "the suite
of 3270 emulators".  Put one README in the base package and you're done.

> I though it necessary to create a separate doc dir & README for each one.

Just don't take the rule too literally.

> See, I'm still working under the assumption that each emulator package is
> independent of the other (expect for -common) and as such should be

But independence of binary packages don't mean they are entirely
independent products.  They share a common purpose and as such they
will be treated by us dumb folks.

Btw., I tried a test of the binary c3270.exe.  From the man page I
got the impression, the emulator should be able to connect to any
telnet server.  I don't have a OS/390 machine handy so I used the
telnet server on Cygwin to connect to.  But it fails.  It connects
and it shows a login prompt.  Then the password is requested by login(1),
the same as running a standard telnet session.  But for some reason
I'm always getting a "Login incorrect" message.  When connecting with
a normal telnet client, I can login, so I don't quite understand how
this is supposed to work.  Any hint?

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.