~ctwm/ctwm/trunk : changes

~ctwm/ctwm/trunk » Changes from revision 719

From Revision 719 to 700

Rev	Summary	Authors	Date
719	Switch to new names for README and CHANGES in installation. Switch to new names for README and CHANGES in installation.	Matthew Fuller	18 days ago
718	Bump cmake minimum version to from 3.6 to 3.22. Bump cmake minimum version to from 3.6 to 3.22. As of 3.31, cmake has deprecated support older than 3.10; while it still works, it complains, and probably won't work for too much longer. Per repology.org, Ubuntu 22.04 includes 3.22, Debian 11 has 3.25 available, CentOS Stream 8 has 3.26, and so on down the line. So this shouldn't be notable to anybody on a reasonably recent system. This allows some minor cleanups, and removal of (probably scarcely tested) conditional support for older versions. Our usage of cmake's vars for finding Xrandr apparently already required 3.14 (addition of X11_Xrandr_FOUND etc), so anybody in the bottom half of this change was already was in bad shape anyway.	Matthew Fuller	18 days ago
717	Doc the SIGCHLD change in changelog. Doc the SIGCHLD change in changelog.	Matthew Fuller	18 days ago
716	Add a meaningless reference to suppress a dead store warning... Add a meaningless reference to suppress a dead store warning from clang's static analyzer.	Matthew Fuller	1 year ago
715	It's a new year and we did things, so bump date. It's a new year and we did things, so bump date.	Matthew Fuller	1 year ago
714	Switch to using a SIGCHLD handler to reap children, instead ... Switch to using a SIGCHLD handler to reap children, instead of SIG_IGN'ing. The intersection of some weird interactions between bits of POSIX and implementation decisions means that on some systems it can cause some unpleasant behavior.	Olaf 'Rhialto' Seibe...	1 year ago
713	Rewrite README and CHANGES into asciidoc. Rewrite README and CHANGES into asciidoc. Previously these were in multimarkdown. While we went straight to asciidoc for the manual, these files require much less careful formatting, so markdown seemed more appropriate. However, they still wanted things like definition lists and some other minor stuff that plain markdown doesn't pull off so well, so the multimarkdown dialect was chosen. However, mmd has proven a little fragile over the years. And the markdown world is so wildly fragmented, and everyone looks at a .md file and says "Oh, that must be processable by my own particular incompatible local dialect", which means anywhere other than our Makefile trying to render the files wound up squirrely anyway. The README is pretty trivially and cleanly asciidoc-able. CHANGES is somewhat less so. We wind up with a lot more noise in it from the heavier markup of asciidoc, so it doesn't read quite as well as plain text anymore. It's probably not too hard to simplify it down to the common subset of markdown. But that would leave it as the _only_ bit of processed markdown in the tree anymore, so I went ahead and adoc-ified it just to be more consistent. So now all our built docs are asciidoc. (n.b. there are a few README.md's in other places through the tree that aren't touched, but they're very much ".txt that happens to look md-ish and may process as that", so what the heck. We have no code for or intention of actually building them into any output formats.)	Matthew Fuller	1 year ago
712	Update astyle config to something current versions accept, a... Update astyle config to something current versions accept, and make indent. This causes a few extra spurious reindents from some astyle changes I can't find or find configs to undo, but they're few enough I'm willing to just eat 'em.	Matthew Fuller	1 year ago
711	Merge release of 4.1.0 and opening up of post-4.1.0 dev. Merge release of 4.1.0 and opening up of post-4.1.0 dev.	Matthew Fuller	2 years ago
710	Remove a testing line accidentally committed in r686 as part... Remove a testing line accidentally committed in r686 as part of the cmake minimum update.	Matthew Fuller	2 years ago
709	Bump copyright date. Bump copyright date.	Matthew Fuller	2 years ago
708	Bypass BorderedLayout for strut'd windows. Bypass BorderedLayout for strut'd windows. This is a hack, and does the wrong thing. However, due to the way we currently handle struts, the previous behavior would force a window to move outside its own strut, which is nonsense. A proper implementation would require evaluating struts individually, not cramming them into BorderedLayout (which already means something similar but distinct), and encapsulating those checks in all the places that do window positioning. That's a major change though; this at least avoids the extra-dumb behavior, and works no worse than 4.0.3 and pre-r644 code for those corner cases.	Matthew Fuller	2 years ago
707	Fixup window placement with DontMoveOff + 3d borders. Fixup window placement with DontMoveOff + 3d borders.	Maxime Soulé	2 years ago
706	Hardcode brz for the version rewrite in tarball generation; ... Hardcode brz for the version rewrite in tarball generation; that's a dev-only thing anyway, and if somebody gen'ing a tarball really needs something else, they'll know how to fix it...	Matthew Fuller	2 years ago
705	We're not expecting any more changes unless we find major bu... We're not expecting any more changes unless we find major bugs, so just call ourselves beta.	Matthew Fuller	2 years ago
704	Merge minor adjustments to shuffle things into a more releas... Merge minor adjustments to shuffle things into a more releasey shape.	Matthew Fuller	2 years ago
703	Make session support built-time conditional. Remains enable... Make session support built-time conditional. Remains enabled by default for now. This allows disabling the X session management protocol support code. This is pretty pointless for the "restart-ctwm" case, since all it's storing is the info the windows and their properties would already have in the first place; it's really only meaningful if you actually use xsm or the like and save and restore full sessions with multiple running clients and their after a system/X restart. It's not clear this was ever a particularly widely used or supported feature across X clients in general, particularly in places where people are using WM's like ctwm as opposed to integrated DE's like Gnome and KDE. It's also a little grody, and for extra points is under slightly different license terms than the rest of the codebase. So, it's possible we may win from retiring it.	Matthew Fuller	2 years ago
702	Hide WindowBox support behind a build-time option, disabled ... Hide WindowBox support behind a build-time option, disabled by default, to be reaped. Some code nearby a few of these ifdefs can probably be slightly simplified once they're gone, but trying to do it conditionally would be too messy to bother with.	Matthew Fuller	2 years ago
701	#ifdef around prototyping captive-related functions. #ifdef around prototyping captive-related functions.	Matthew Fuller	2 years ago
700	Bump copyright date. Bump copyright date.	Matthew Fuller	2 years ago

Older »