Re: [ITP] trie-gen 1.1

[Christopher Faylor <cgf at redhat dot com>]

On Tue, Sep 23, 2003 at 11:49:12PM -0400, Igor Pechtchanski wrote:
>On Tue, 23 Sep 2003, Christopher Faylor wrote:
>
>> On Tue, Sep 23, 2003 at 06:21:24PM -0400, Igor Pechtchanski wrote:
>> >I've just managed to build trie-gen (part of libg++) for separate
>> >packaging.  Trie-gen generates a minimum-prefix trie for a set of keywords
>> >(essentially a trie counterpart for gperf).
>>
>> This can't be a coincidence, right?  I just asked for something like this
>> over in cygwin-developers.
>>
>> Did you find this through google?  libg++ is pretty ancient.
>> cgf
>
>Well, sort of.  I suggested tries in reply to Corinna's message just
>before I got yours, and a Google search for "gperf" turned up "trie-gen"
>for me.  Finding a reasonably recent source was quite another story.

Huh.  I somehow missed your message in cygwin-developers.  I just checked
now and there it was.  Sorry for the confusion.

>Does anyone know of a newer version?  Maybe even one that's still under
>development?

I'm using tries for the mount table in another ancient sandbox, so I've
always had tries in the back of my mind for this application in cygwin.
My tenth google search unearthed the 'shilka' package which seems to do
what I want.  I've got my branch built using that package so I may just
switch to it.

http://cocom.sourceforge.net/shilka.html

While I know some of the guys who wrote trie-gen, I also know and respect
the person who wrote shilka and, in fact, he'll be working for me at
Red Hat next week, oddly enough.  Shilka builds out of the box on linux
and I suspect it will build fine on cygwin, too.  Would you be interested
in packaging the cocom package (of which shilka is a part) for cygwin?

cgf