Updates for stuff mostly discussed in bug #12232 New ebuilds follow for: * dev-libs/libax25 * media-radio/ax25-apps * media-radio/ax25-tools
Created attachment 14148 [details] dev-libs/libax25-0.0.11-r1 canidate dev-libs/libax25 (ebuild by Robert Steinh
Created attachment 14148 [details] dev-libs/libax25-0.0.11-r1 canidate dev-libs/libax25 (ebuild by Robert Steinhäußer / gentoo _at_ robsh.de): - Cleanup by Robert Steinhäußer of the libax25 script. - Fixes license error noted in bug #12232 but never fixed - Adds user documentation (dodoc statement) - Installs all the sample files (axports, nsports, rsports), as opposed to providing a Gentoo-ized version of just axports(*) (*) This I am going to argue over a bit. I understand I forgot nsports and rsports sample configurations in my previous ebuild. But I would prefer that Gentoo not ship with their default values being a legitimate series of call signs. In my view they should be N0CALL (obviously bogus), and commented out. But that's just me being legally paranoid.
Created attachment 14149 [details] media-radio/ax25-apps-0.0.6-r1.ebuild canidate media-radio/ax25-apps-0.0.6-r1.ebuild candidate (by Samuel Greenfeld): - Slight cleanup (default style src_compile removed) - SRC_URI now in "mirror://" format (like new libax25 one) - Documentation forcibly installed in a gentoo-like manner (to /usr/share/doc/<ebuild_name_and_revision>) - A method to protect an existing /var/lib/ax25/ax25rtd is _still_ required. One can detect if this directory already exists, but I do not know if it is desirable to set CONFIG_PROTECT in the middle of an ebuild. While we can simply not overwrite the existing database if we see the directory exists, we do not want to remove it if it still contains our blank template files. But we may have to remove files from this directory on a version update. More research required (I need to stick one of my TNCs into KISS mode and collect packets for a while). Perhaps this functionality should be moved to pkg_postinst()? - This ebuild _still_ requires the .rc init scripts previously imported into the Portage media-radio/ax25-apps tree. - I suspect the "no signs of failure" problem with daemon startups still exists.
- Specifically: ax25ipd and ax25rtd report "OK" back to the init scripts, but require zapping if their post-fork routines fail. An easy way to force this is to disable all your AX.25 Linux interfaces. I'm looking at the programs' C code, but I see no easy obvious fix (maybe opening the sockets before the fork() with ax25rtd, but I'm not sure if that's allowed). Ax25mond doesn't appear to have this issue. But if all your AX.25 interfaces are down, it's monitoring nothing.
Created attachment 14152 [details] media-radio/ax25-tools-0.0.8-r1.ebuild candidate (By S. Greenfeld) Derived from the "First cleanup" variant found in bug #12232, which actually is a bit newer than what is in the current portage tree. - Conditional X Windows detection. Really just controls if the soundmodem diagnostics program (smdiag) is made or not IIRC. - Program name correction in the ebuild (versus current portage variant). - More Gentoo-styled documentation installation. - Daemon startup files for things that have configuration files in /etc/ax25 (ax25d, netromd, rip98d, rxecho, ttylinkd). - Daemon startup files for a daemon that lacks configuration files and requires very little configuration (mheardd). - A lot of stuff in this package (kissattach, etc.) likely should be run before the above daemons start. Does anyone have a proposal for a nice, generic way to do this (regardless of using kiss/mkiss/rose/netrom/sm/etc.)? - Some testing done on my dual-port TNC (with mkiss/kissattach/mheard/etc.), but I'd have to find someone to peer with to do Netrom/etc. testing. Rose testing should be done by someone in a Rose-utilizing area.
Samuel... Do you have the geo_trans.diff? The bug tracker is missing the file.
my mistake on the blocker. removed
committed... nice work.
