Re: UPDATE: rxvt-unicode - 9.16
On Fri, Dec 28, 2012 at 10:13 PM, Christian Weisgerber na...@mips.inka.de wrote: David Coppa dco...@gmail.com wrote: Seems fine to me so far (amd64 and macppc); OK here - amd64. Ok for me too. CONFIGURE_ARGS =--enable-utmp From this I take it that rxvt is supposed to write entries to utmp(5), visible with w(1). And as an xterm replacement it should. In reality, it doesn't write utmp entries. It can't. It would need to be setgid utmp for this. The code seems to be there, so I simply made it setgid utmp. Surprisingly, this causes urxvt to fail completely: urxvt: protocol error while reading pty/tty fds from helper process, aborting. I've looked a bit over the code and ktrace'd a run, but there is no obvious reason for the error. I suspect there is a problem with the IPC, probably the file descriptor passing. I'll have a look, thanks for reporting this problem. Btw, '--enable-utmp' and '--enable-wtmp' were *not* in the original patch Levai sent to ports@. ciao, David
Re: UPDATE: rxvt-unicode - 9.16
On Fri, Dec 28, 2012 at 2:10 AM, Aaron Bieber def...@gmail.com wrote: On Thu, Dec 27, 2012 at 10:23:34PM +, Stuart Henderson wrote: On 2012/12/27 22:35, LEVAI Daniel wrote: Hi! Update for rxvt-unicode. Use patch -E, one file from patches/ is gone. Also regenerated WANTLIBs. A few changes worth noting (at least in my opinion) from the changelog: - perl extensions now have their own manpages, installed as urxvt-NAME(1) - completely redone background pixmap support - users can now specify arbitrary expressions (And programs) that calculate new terminal backgrounds. - perl extensions can now provide commandline switches and resources which show up in -help etc. - remove support for libafterimage. (we didn't use it, however) - update to CVS version of libev Seems fine to me so far (amd64 and macppc); I'm OK with it but would like to hear from one or two other people who use this before committing. OK here - amd64. Ok for me too. I'll also sync devel/libev with upstream. Thanks, David
Re: UPDATE: rxvt-unicode - 9.16
David Coppa dco...@gmail.com wrote: Seems fine to me so far (amd64 and macppc); OK here - amd64. Ok for me too. CONFIGURE_ARGS =--enable-utmp From this I take it that rxvt is supposed to write entries to utmp(5), visible with w(1). And as an xterm replacement it should. In reality, it doesn't write utmp entries. It can't. It would need to be setgid utmp for this. The code seems to be there, so I simply made it setgid utmp. Surprisingly, this causes urxvt to fail completely: urxvt: protocol error while reading pty/tty fds from helper process, aborting. I've looked a bit over the code and ktrace'd a run, but there is no obvious reason for the error. I suspect there is a problem with the IPC, probably the file descriptor passing. -- Christian naddy Weisgerber na...@mips.inka.de
Re: UPDATE: rxvt-unicode - 9.16
On 2012/12/27 22:35, LEVAI Daniel wrote: Hi! Update for rxvt-unicode. Use patch -E, one file from patches/ is gone. Also regenerated WANTLIBs. A few changes worth noting (at least in my opinion) from the changelog: - perl extensions now have their own manpages, installed as urxvt-NAME(1) - completely redone background pixmap support - users can now specify arbitrary expressions (And programs) that calculate new terminal backgrounds. - perl extensions can now provide commandline switches and resources which show up in -help etc. - remove support for libafterimage. (we didn't use it, however) - update to CVS version of libev Seems fine to me so far (amd64 and macppc); I'm OK with it but would like to hear from one or two other people who use this before committing.
Re: UPDATE: rxvt-unicode - 9.16
On Thu, Dec 27, 2012 at 10:23:34PM +, Stuart Henderson wrote: On 2012/12/27 22:35, LEVAI Daniel wrote: Hi! Update for rxvt-unicode. Use patch -E, one file from patches/ is gone. Also regenerated WANTLIBs. A few changes worth noting (at least in my opinion) from the changelog: - perl extensions now have their own manpages, installed as urxvt-NAME(1) - completely redone background pixmap support - users can now specify arbitrary expressions (And programs) that calculate new terminal backgrounds. - perl extensions can now provide commandline switches and resources which show up in -help etc. - remove support for libafterimage. (we didn't use it, however) - update to CVS version of libev Seems fine to me so far (amd64 and macppc); I'm OK with it but would like to hear from one or two other people who use this before committing. OK here - amd64.