This is the mail archive of the mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

To ITP or not to ITP, that is the question...

(With apologies... :)

I have a quandry.

In the process of updating some packages I maintain for Cygwin (zsh &
suite3270), I've found that I now require some other software to build
them.  One is 'ICU' (International Components for Unicode) and the other
is 'yodl' (Yet oneOther Document Language).  I need yodl to generate the
doc for the developer snapshots of zsh (which I'm trying to create any
automated process for), and I need ICU for DBCS support in suite3270.

I've already build both of the above packages (in fact, I've gone the
next step and created Cygwin packages for them :).  I just don't know
weither there is any interest in these being offered as Cygwin packages.

So, my quandry is: Should I offer them?  Is there any interest from other
Cygwin users?  It's not a question of maintaining them, really.  Since I
now require them for other packages I maintain, I have to keep my private
copies up to date anyways.  And, since I've already gone through the
exercise of creating packages, there's not much left to do.

You can check them out at



There are two wrinkles:

1) There's some confusion in my mind about yodl's viability.  I've seen
notices on the web about it being discontinued, yet the sourceforge
project was last updated in Jan of 2005.  I've build v1.31.18 so far. I
haven't gotten around to build v2.01.01 yet (it requirs 'icmake'!).  I
guess I'll have to look into building icmake next :|

2) ICU is an IBM product (at 9Mb, compress, for the source, it's hugh!).
Normally that's got both good and bad connotations.  This particular
release of ICU (3.2) is strange in that I can't seem to get the canned
sample demos to work correctly under Cygwin.  They run, but don't
generate the desired output.  That normally would bother me, except, a
build of the same version under Linux also has the sample demos failing
in the same way.  Well, at least it's consistent :) But, it begs the
question: does this code actually work at all?  I'd presume IBM thinks
so, otherwise why would they release it?  So, even though I've managed to
build and package it, it could be complete junk.

*sigh* It's never easy, is it? Please send me your thoughts on this.

Peter A. Castro <> or <>
	"Cats are just autistic Dogs" -- Dr. Tony Attwood

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]