On Sat, May 10, 2014 at 02:58:17AM +0100 I heard the voice of Aaron Sloman, and lo! it spake thus: > > FIRST PROBLEM: > missing files in /usr/include/gnu > stubs-64.h > stubs-32.h > > solved by tip found using google: > yum install glibc-devel.x86_64
These aren't headers we use ourselves; that's something internally used by other system headers. <http://forums.fedoraforum.org/showthread.php?t=271558> suggests they're actually needed by stdio.h (through a chain of other headers it uses). So it's a long way down the chain from anything we directly use. I don't think we could reasonably test it any way other than the way we do, which is blowing up trying to compile stuff :) > SECOND PROBLEM: [...] > Solved by > yum install libXmu-devel > Similar problem solved by > yum install libXpm-devel Yeah, these are standard bits in there. And mostly stuff unchanged since forever; e.g., that X11/Xmu/CharSet.h include has been there since at least ctwm-1.1. Same in current twm, so that probably predates even Claude. So this is probably "first time I compiled ctwm" surprises rather than "new in ctwm 3.8.2" surprises. (don't get me started on Linux and their freakin' "Oh, I see you installed this library, that must mean you never want to ever compile anything using it, so we don't need to install header files" fetish. Whoops, I just did...) > I inserted: > > #include <stdint.h> > > Then 'make' completed, somewhat to my surprise! Should be fixed upstream now for the future. > I am left wondering where all the information about workspaces, > mappings of windows to workspaces, locations on the screen, whether > minimised or not, etc. was stored while ctwm was not running. I had > assumed the data-structures were maintained by ctwm. Most of it is actually stashed in X properties, at least the ones that have standard such defined. You can see those with the xprop(1) command. For instance, looking at the xterm I use for IRC, it has stuff like WM_NORMAL_HINTS(WM_SIZE_HINTS): user specified location: 0, 532 user specified size: 499 by 329 program specified size: 499 by 329 program specified minimum size: 25 by 17 program specified resize increment: 6 by 13 program specified base size: 19 by 4 window gravity: NorthWest It's on the left side of my screen, in the lower half, which the location covers. Apparently the font glyph size is 6x13, so it's hinting to the WM that it should be resized in steps that size (which I don't think ctwm listens to). WM_OCCUPATION(STRING) = "One", "Two", "Three", "Four", "Six", "Seven", "Eight", "Nine" I've got 9 workspaces, with #5 kept empty, and IRC is one of those things I want in all of them, so it's got 'em all there. > Presumably all this means it should be straightforward for someone > to build fedora rpm files. Is the previous fedora rpm builder on > this mailing list? Going through http://pkgs.org/search/?query=ctwm&type=smart leads to http://pkgs.org/fedora-20/rpm-sphere/ctwm-3.8.1-3.3.x86_64.rpm.html for a "current" RPM. Last few people listed in the Changelog aren't people I recall ever seeing around the list. The lines from 3.8a earlier list Richard, but I'm not sure he ever actually made or uploaded the RPM's; those are from the .spec file in the source tree. Seems like the last 2 packagings use forked .spec's. So maybe that someone can be you :) On the Debian side, Branden seems no longer involved (and his list mails bounced, so he's apparently no longer around here either); that pkgs.org search only finds 3.7 .deb's. Though there are NMU's up through earlier this year (changing build/package stuff, seemingly not code). -- Matthew Fuller (MF4839) | [email protected] Systems/Network Administrator | http://www.over-yonder.net/~fullermd/ On the Internet, nobody can hear you scream.
