Date: Wed, 29 Mar 1995 21:10:26 +0200 (MSZ) From: me@dude.pcs.dec.com ( Michael Elbel ) To: asami@cs.berkeley.edu (Satoshi Asami/=?ISO-2022-JP?B?GyRCQHUbKEI=?= =?ISO-2022-JP?B?GyRCOCsbKEIgGyRCOC0bKEI=?=) Cc: me@FreeBSD.org, ports@FreeBSD.org Subject: Re: libXpm.so.4.5 Message-ID: <m0ru38P-0005fOC@dude.pcs.dec.com> In-Reply-To: <199503291438.GAA04782@silvia.HIP.Berkeley.EDU> from "Satoshi Asami/=?ISO-2022-JP?B?GyRCQHUbKEI=?= =?ISO-2022-JP?B?GyRCOCsbKEIgGyRCOC0bKEI=?=" at Mar 29, 95 06:38:09 am
next in thread | previous in thread | raw e-mail | index | archive | help
> The current numbering scheme is correct. The "3" is the "protocal > version" or something, and the "4" and the alphabet denote the major > and minor shlib version numbers. Thus, 3.4e -> Xpm.4.5. Aaaaah, that actually makes sense, doesn't it? :-) > However, the original port was done incorrectly by someone who thought > 3.4c corresponds to libXpm.so.3.4, and I believe there was a bogus > #define in one of the patches that rewrote 4.3 -> 3.4. I see. This means we should be safe now until xpm-3.5 comes out but that this then would mean work on the clients that use it anyways, right? > Yikes. When we discovered the problem, I went to thud to do a ldd in > /usr/local/bin and /usr/X11R6/bin and remade all the packages that had > Xpm.3 in them. Didn't check all the packages though. :< Would have been quite some hassle too - untarring all the packages and ldd'ing the binaries. Maybe we should make it a rule that each and every port should be installed on thud. Then clean them out maybe once a month and try to rebuild and reinstall them. Then make new packages. > Thanks for your help. For fvwm, do you mean we can't replace the > version in package/ without introducing a new problem by compiling now, No, recompiling fvwm will generate a package identical to what is in packages now (except for it using libXpm.so.4.*). In any case we should be careful about generating packages that go into the ftp area on -current machines. The libc minor version bump will hit everybody who's just installed the latest SNAP and wants to get some packages for his machine (something like "I want libc-2.1 but can only find libc-2.0, I'm using it anyways) watch out for those bug reports :-(. > or do you mean it's already broken about the .xpm stuff in the > package that we currently have? If it's the latter, I'll just rebuild > fvwm and replace it too for now. Right, the current package is at least incomplete. I've just tried it (rebuilding on -current). We've installed the 0322 snap on a virgin box here and wanted to make a decent workstation out of it using the packages (and pkg_manage - nice interface, congratulations). Doing this I discovered that running a just pkg_added fvwm won't work. The GoodStuff module will complain about not being able to find its pixmaps, which wouldn't be so bad. Unfortunately it seems that the main fvwm also relies on the pixmaps. It, however, does *not* complain but simply refuses to display the titlebar buttons and react to mouse clicks. I consider this seriously broken ;-) Then there's the issue of autoraise not working (it will give the focus to the window the mouse cursor is in but not raise the window). Then there's the thing with fvwm wanting to find its files under /usr/lib/X11 and /usr/include/X11. Those paths should be symlinks to /usr/X11R6/lib/X11 and /usr/X11R6/include (and there should be a /usr/X11R6/bin -> /usr/bin/X11). I *think* this should be done by the XFree86 install script. Fixing the Makefile to copy the .xpm files and add them to the package will be trivial. It's the other two points that need to be addressed too at least for the 2.1 CD. I suggest removing the fvwm package for now until it works. I'm willing to look into it but can't promise to get this done before the end of next week. Btw. I'm baffled by the number of ports you manage to handle. A quck grep finds 'asami' in the $Id$ strings of 64 makefiles. The ports collection would be a lot less attractive without your work :-) Michael -- Michael Elbel, Digital-PCS GmbH, Muenchen, Germany - me@FreeBSD.org Fermentation fault (coors dumped)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?m0ru38P-0005fOC>