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: cygipc (and PostgreSQL) XP problem resolved!

Robert Collins wrote:

key_t, as it's used for ipc, is likely to be *problematic* to transition
in a 'fat binary' style.

You'd need a 32 bit set of key creation routines, and and translation
table to lookup 32bit keys in the list of 64 bit keys ....

Which basically aliases the entire 64bit key space down to 32bit space -- which kinda short circuits whole reason that cygdaemon wanted 64bits in the first place.

I don't think it's worthwhile to do a 'fat binary' style implementation for key_t.

Given that cygipc is *not* in cygwin today, and you'd be adding it, I'd
simply have it 64 bit from the first release uploaded to sources.

Yes, I agree -- but (obviously) only if newlib/cygwin decide that the 64bit key_t definition is a good idea, and accept a patch to do so.

And I'd time that for oh, a day after cygwin 1.5 goes up as a testing
package. (And release cygipc as testing whilst cygwin 1.5 stays in

Yeah, that sounds reasonable.

I take it you're in favor of adding cygipc to the distro (or are you speaking academically)?


Unsubscribe info:
Problem reports:

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