You're absolutely right. I should have thought before I posted. In fact I already patched ibcs to work regardless which gcc/C library is used, since ibcs (as a kernel module) should include the kernel headers, but not the differing application headers.
I'm sorry for reporting a non existant bug. Michael P.S.: It was one of those days ... -- Dr. Michael Meskes, Projekt-Manager | topystem Systemhaus GmbH [EMAIL PROTECTED] | Europark A2, Adenauerstr. 20 [EMAIL PROTECTED] | 52146 Wuerselen Go SF49ers! Use Debian GNU/Linux! | Tel: (+49) 2405/4670-44 >---------- >Von: Helmut Geyer[SMTP:[EMAIL PROTECTED] >Gesendet: Freitag, 23. Mai 1997 19:48 >An: [EMAIL PROTECTED]; [EMAIL PROTECTED] >Cc: [EMAIL PROTECTED] >Betreff: Re: Bug#10060: libc6-dev: time.h and linux/time.h conflict > >> Package: libc6-dev >> Version: 2.0.3-4 >> >> Just try to compile ibcs. I won't work because one source file includes >> <time.h> and <linux/sched.h> (or others), which includes <linux/time.h>. >> This gives an error condition since both files have a different definition >>of >> structure timespec. In fact it's not that much of a difference (long vs. >> long int) but enough for gcc to see an error. There are other double >>defined >> structures in other files (e.g. timeval, timezone, itimerval). >> >That is right, but this is _not_ a bug in libc 6 but comes from the fact that >many Linux applications use the kernel haeders themselves, which generally is >a bad idea anyway. >The change in the C library (i.e. to provide headers for all the stuff that >is >not really kernel-dependant) is coordinated with Linus and the kernel >developers. >These problems are bugs (or lets say portability problems) in the packages.. >IBCS should be fixed, not libc6. > > Helmut > >-- >Helmut Geyer >[EMAIL PROTECTED] >public PGP key available : finger >[EMAIL PROTECTED] > -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
