A patch for the netatalk-2.0.3-r2 ebuild that adds support for zeroconf ( through avahi, howl and mdnsresponder )
Created attachment 86910 [details, diff] netatalk-2.0.3-afpd-zeroconf.patch (goes in the files folder!)
Created attachment 86911 [details] netatalk-2.0.3-r3.ebuild
Comment on attachment 86910 [details, diff] netatalk-2.0.3-afpd-zeroconf.patch Adds zeroconf support to netatalk
Comment on attachment 86911 [details] netatalk-2.0.3-r3.ebuild Adds zeroconf support for netatalk
This is a quite intrusive patch, is that applied upstream? I'm quite interested in this, but it might be a bit overweight on maintenance if it's not upstreamed. If it's applied in upstream CVS I'd rather take a snapshot of it.
Original patch from which this ebuild / patch is based can be found here: https://sourceforge.net/tracker/index.php?func=detail&aid=1474791&group_id=8642&atid=308642 Upstream development appears to be fairly slow. Whist I appreciate your reluctance to apply this patch we might be waiting a while. (6 months since the last checkin).
Created attachment 87045 [details] netatalk-2.0.3-r3.ebuild Fixes an issue with the init file (now ensures that the respective zeroconf daemon is running).
hi can't compile on my amd64. the error is cnid_index.c: In function 'dbif_count': cnid_index.c:277: warning: passing argument 2 of 'db->stat' from incompatible pointer type cnid_index.c:277: error: too few arguments to function 'db->stat' make[3]: *** [cnid_index.o] Error 1
Created attachment 139907 [details] Fixed regression that makes the build break with more modern versions of DB4 Looks like the previous ebuild script had a regression which causes the build to fail when a more modern version of db is used. This should fix the regression ..
The patch netatalk-2.0.3-afpd-zeroconf.patch is faulty. afp_zeroconf.h:40:8: warning: extra tokens at end of #endif directive caused by uncommented stuff after the #endif afp_bonjour.c: In function 'bo_zeroconf_setup': afp_bonjour.c:45: error: expected expression before ',' token caused by a double comma (probably a typo) at line 45 of afp_bonjour.c afp_bonjour.o: In function `bo_zeroconf_unregister': afp_bonjour.c:(.text+0x23): undefined reference to `DNSServiceRefDeallocate' afp_bonjour.o: In function `bo_zeroconf_run': afp_bonjour.c:(.text+0x88): undefined reference to `DNSServiceRefSockFD' afp_bonjour.c:(.text+0xbc): undefined reference to `DNSServiceProcessResult' afp_bonjour.o: In function `bo_zeroconf_setup': afp_bonjour.c:(.text+0x1a1): undefined reference to `DNSServiceRegister' collect2: ld returned 1 exit status make[3]: *** [afpd] Error 1 After this i stopped bothering with it.
upstream has issues with the patch, and it needs updating to netatalk-2.0.5 ...