This is the mail archive of the
cygwin@cygwin.com
mailing list for the Cygwin project.
RE: MIT shared memory extension
- From: "Ralf Habacker" <Ralf dot Habacker at freenet dot de>
- To: "Robert Collins" <robert dot collins at itdomain dot com dot au>, "cygwin" <cygwin at cygwin dot com>
- Date: Fri, 17 May 2002 13:58:18 +0200
- Subject: RE: MIT shared memory extension
> > Thats all. No patching headers. This enables migrating one
> > package to cygwin ipc stuff, while other packages could use
> > the cygipc stuff.
>
> We still need to do the ABI upgrade though, before that can be done.
You mean building a new cygwin1.dll with shm support ?
I have patched sys/types.h key_t definition with
#ifdef _CYGIPC_IPC_
typedef long key_t;
#else
typedef long long key_t;
#endif
have applyed the cygwin.din patches, have recompiled and installed cygwin1.dll
in /usr and than I can compile an app using cygipc and another using cygserver.
What's additional to do ?
So the only thing to do I see is to distribute a new cygwin release with shm
support.
>
> > BTW: This is another thread, but perhaps it is interesting
> > for somebody to see. Below there is a first performance
> > measurement of cygipc relating to cygserver shm support. Only
> > for info.
>
> Interesting. How to read the results? (is lower better or worse?).
>
This results are produced by the profiler lib located in the
http://kde-cygwin.sf.net cvs area
n = nummer of running tests
min = minimal measured time for one test
max = maximal measured time for one test
avg = average time of all tests
sum = accumulated time for all tests
the results are in usec
> Also, as you can read both codes without issue, would you care to
> document how the cygipc one is much faster?
Hmmh, I have only some assumption, but I will think about this.
BTW: What about a new binutils release. Is this going on ?
Regards
Ralf
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/