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]

Re: minires-0.95 - a new package ready for review

At 10:33 PM 6/11/2003 -0400, Nicholas Wourms wrote:
>Is there any reason why this cannot be put in the cygwin (or newlib) 
>source tree?  AFAICT, Pierre used BSD licensed source code as a basis 
>for his work, so it wouldn't violate RedHat's licensing if Pierre gave 
>his ok.  This is definitely an area where newlib is deficient (unless 
>you are linux-based) and thus impedes the ability for portability from 
>*nix -> win.  This is basic core functionality nowadays, and as such it 
>ought to be included with the core distribution.  At the very least, it 
>is installing POSIX system headers, so that ought to be justification 
>enough to include it.  We already have started (obviously not completed) 
>doing this in another case, IPC semaphores.  It seems that similar 
>reasoning should apply to the resolver services.  Now as to whether it 
>should be spun as a separate dll/import library (-lresolv), is another 
>matter, entirely.  Of course, feel free to disagree, as I could be wrong 

Yep, we had this discussion when I proposed exim, 15 months ago. There 
was no consensus at all (minires in cygwin, minires as a dll, a real
bind package, something else in cygwin... ) so I ended up with a direct
compile in exim. 

Several projects have used minires, without apparent problems.
It's used mostly by small group of developers who find the source
by searching the cygwin list or googling, so I didn't push for a
package. There is also the issue of potential system header conflicts.

At any rate, thanks to Jari Aalto for the packaging effort.


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