Re: Ctrl-C propagates
On Friday 10 October 2003 21:46, Harold L Hunt II wrote: Dia, Hmm... I don't understand what is happening here. Can anyone else confirm this? Yes. Scenario here (Win2k Pro V5 Build 2195 SP4) for me is: cyg startx from a bash console (fires a window manager-less xterm). cyg ssh -X [EMAIL PROTECTED] (me FreeBSD 5.2 box) bsd wmaker A ctrl-c at the cyg bash console kills all X processes on bsd. Ditto if I fire up kde instead of windowmaker (and ditto kde/gnome apps from inside windowmaker). The X server is left running - appears to be the only thing immune to the ctrl-c. Quit it windoze style (click closer) and we're back to original cyg bash console. Aside from the uncerimonious terminations there don't appear to be any side-effects. If one were to imagine the ctrl-c being broadcast to all X descendants that about sums it up. $ cygcheck -s |grep ssh\|XFree\|cygwin1 1083k 2004/01/31 E:\cygwin\bin\cygwin1.dll DLL identifier: cygwin1 Shared id: cygwin1S3 openssh 3.7.1p2-2 XFree86-base4.3.0-1 XFree86-bin 4.3.0-8 XFree86-bin-icons 4.3.0-5 XFree86-doc 4.3.0-1 XFree86-etc 4.3.0-6 XFree86-f1004.2.0-3 XFree86-fenc4.2.0-3 XFree86-fnts4.2.0-3 XFree86-fscl4.2.0-3 XFree86-fsrv4.3.0-7 XFree86-html4.3.0-1 XFree86-lib 4.3.0-1 XFree86-lib-compat 4.3.0-2 XFree86-man 4.3.0-2 XFree86-nest4.3.0-5 XFree86-prog4.3.0-12 XFree86-prt 4.3.0-5 XFree86-ps 4.3.0-1 XFree86-startup-scripts 4.2.0-5 XFree86-vfb 4.3.0-5 XFree86-xserv 4.3.0-44 I can do some tests if prompted. -- Guy Harrison [EMAIL PROTECTED]
Re: building XFree86 from cvs
On Tue, 27 Aug 2002 11:14:56 -0700 (PDT), Nicholas Wourms [EMAIL PROTECTED] wrote: --- Guy Harrison [EMAIL PROTECTED] wrote: On Tue, 20 Aug 2002 10:13:50 -0500, Michael Harnois [EMAIL PROTECTED] wrote: Oooh. This is stranger than I thought. Some of those programs do get built correctly, despite the log messages. For instance in config/util, makestrs, revpath, and rman build ... but lndir doesn't. Probably because gcc -o foo yields foo.exe automatically but unfortunately dependencies and so forth won't know that. Your average configure make app can sometimes get away with it - only to fail on the install uninstall (good indication of this problem existing is make invokes the linker even though the last make succeeded). Sadly this is my first forray into the world of Imake so take my advice for this particular problem with a strong pinch of salt. You'd think TOG would get a clue and switch to the autotools... All that's required is for me to become 100% proficient in Imake. It would guarantee it! :-) -- [EMAIL PROTECTED]
Re: building XFree86 from cvs
On Mon, 26 Aug 2002 07:11:37 -0400, Harold Hunt [EMAIL PROTECTED] wrote: Guy, Okay, this looks reasonable. I will take a look at it and submit it when I get a chance. Classes start today, so it may be a few days. No problem. Things progress slow my end! -- [EMAIL PROTECTED]
Re: building XFree86 from cvs
On Thu, 29 Aug 2002 22:06:58 +0200 (MEST), Alexander Gottwald [EMAIL PROTECTED] wrote: On Sun, 25 Aug 2002, Guy Harrison wrote: [big diff] ...it might even still cross-compile! ;-) It does. Good Work :) Thanks! bye ago BTW: When you're creating diffs, please use the diff -u. Noted. -- [EMAIL PROTECTED]
Re: building XFree86 from cvs
On Fri, 23 Aug 2002 17:48:05 -0400 (EDT), Harold Leatherman Hunt II [EMAIL PROTECTED] wrote: Well, I can't offer any help because I have no idea what you are doing. In order for me to help, or for me to even investigate this problem any further, you are going to have to follow, exactly, the instructions for compiling that are in the Cygwin/XFree86 Contributor's Guide (http://xfree86.cygwin.com/docs/cg/). I have, except for the folders. I'm compiling into obj-dbg and obj-rel for debug and release versions respectively. Doesn't matter whether the build is (makeg.sh)debug or release: failures are at the same points. You will then have to be very clear about whether you are cross compiling or whether you are building on Cygwin, Building on Cygwin. NT Workstation SP6a RC1.28. CYGWIN=binmode ntsec codepage:oem and you will have to tell me the precise step in the Contributor's Guide instructions that you start to have problems at. You will also have to send in your World.log file (which will be generated if you follow the CG instructions). Step 7. make World which for the most part works. It's simply a few .exe extensions getting added which are causing make's dependencies to fail. Typically foo.exe.exe but sometimes foo.exe.blah. As Alexander suggested, I performed (my first ever!) diff of my Imakefile alterations against CVS tonight. It ought to precisely identify the problem even if I've made a complete mess of doing them in the required format. I simply cannot spend time debugging each developers command-sequence that they use for building. You've got to stick to the prescribed sequence in order to get help from me. Fair enough. -- [EMAIL PROTECTED]