~ctwm/ctwm/trunk : changes

~ctwm/ctwm/trunk » Changes from revision 604.1.7

From Revision 604.1.7 to 592

Rev	Summary	Authors	Tags	Date
604.1.7	These elements are also used in Pixel fashion, so make 'em s... These elements are also used in Pixel fashion, so make 'em so.	Matthew Fuller		7 years ago
604.1.6	Now that these docs are short, collapse 'em to ///< at EOL t... Now that these docs are short, collapse 'em to ///< at EOL to save a little space.	Matthew Fuller		7 years ago
604.1.5	These are everywhere used as Pixel's, so might as well decla... These are everywhere used as Pixel's, so might as well declare 'em as such.	Matthew Fuller		7 years ago
604.1.4	Add braces around this if to fit our style. Add braces around this if to fit our style.	Matthew Fuller		7 years ago
604.1.3	Eliminate ScreenInfo.SqueezeTitleSet. It's used as a flag i... Eliminate ScreenInfo.SqueezeTitleSet. It's used as a flag in the config file parsing, for the sole purpose of avoiding a duplicative store by doing a comparison. I'm not sure that's actually a win in CPU terms anyway, never mine the loss in readability and waste of space in the struct the rest of runtime.	Matthew Fuller		7 years ago
604.1.2	Both sides of this conditional are identical, so why have it... Both sides of this conditional are identical, so why have it?	Matthew Fuller		7 years ago
604.1.1	This var is used as a bool, so make it one. This var is used as a bool, so make it one.	Matthew Fuller		7 years ago
604	astyle won't get along with aligning these multi-line ///< c... astyle won't get along with aligning these multi-line ///< constructs in the doxygen comments in screen.h. Convert them all to preceeding /// style instead, and update our devman mentions to explain why.	Matthew Fuller		7 years ago
603	Add DontNameDecorations config var to disable the XStoreName... Add DontNameDecorations config var to disable the XStoreName()'s on our wrapper/decoration windows.	Matthew Fuller		7 years ago
602	Fix up a typo I introduced into the config parsing before 4.... Fix up a typo I introduced into the config parsing before 4.0.0.	Matthew Fuller		7 years ago
601	Document up all the fields in the ScreenInfo structure, and ... Document up all the fields in the ScreenInfo structure, and most of the rest of screen.h. This is all done in Doxygen-style markup, so tweak the doxygen build system pieces a bit, and add a devman doc about it, now that it's useful for something. Some fields had existing documentation that was just converted into doxygen comments. Many others were previously undocumented and required figuring out and documenting. In a number of cases, the existing documentation was just truncated to the name of the config var they came from, since the meaning is either pretty obvious from the name, or sufficiently complex that we'd be rewriting the manual in comments.	Matthew Fuller		7 years ago
600	Do commenting and minor code cleanup through the KeyPress ha... Do commenting and minor code cleanup through the KeyPress handler as part of investigating some menu handling issues. Should be no functional change. Some small comments also added to other places involving opening/invoking menus, to be sure it's documented what magic happens in them.	Matthew Fuller		7 years ago
599	Correct speeling: beeing -> being. Correct speeling: beeing -> being.	Matthew Fuller		7 years ago
598	Minor cleanups in m4 parsing code. No user-visible changes. Minor cleanups in m4 parsing code. No user-visible changes. - Remove MkDef/MkNum functions and just use fprintf(). - Allocate space on the heap for the temp file name and honor $TMPDIR while we're at it. - Some variable scope-narrowing and rewrites away from fixed-size buffers.	Matthew Fuller		7 years ago
597	Patch over a bug in generating the temp file for m4 stuff. Patch over a bug in generating the temp file for m4 stuff. When multiple X Screen's are used, we run through building the m4 config file bits for each, since some of the defs change. That means multiple runs through m4_defs() to build the file, which means multiple mkstemp() calls to build the temp file. But since tmp_name is static, the second time through it would already have filled out the temporary name. This is sorta safe absent malice, since we'd already have a safe file the first time around, so it would take somebody trying to race with us to make us overwrite something. But perhaps not entirely, and malice is out there. And some implementations of mkstemp() (e.g., glibc) will error out if the template filename doesn't contain an actual template, which means running with multiple Screens on those systems will actually lead to ctwm perror()'ing. Since it's returning that var, we can't just make it non-static, so re-initialize it with the string each time through. A larger solution should probably restructure things a little around this code... See also: https://bugs.gentoo.org/show_bug.cgi?id=621024	Matthew Fuller		8 years ago
596	Bump version to post and prepare for $NEXT. Bump version to post and prepare for $NEXT.	Matthew Fuller		8 years ago
595	Bump version and cut 4.0.1 with the OTP fixes. Bump version and cut 4.0.1 with the OTP fixes.	Matthew Fuller	ctwm-4.0.1	8 years ago
594	Add a guardrail so removing an EWMH _FULLSCREEN state on a w... Add a guardrail so removing an EWMH _FULLSCREEN state on a window that's never been zoomed around (_just_ possibly triggerable in other esoteric ways, though I don't know them) doesn't resize it down to nothing.	Matthew Fuller		8 years ago
593	Merge fix for stacking problem that leads to OTP assertion f... Merge fix for stacking problem that leads to OTP assertion failures and aborts.	Matthew Fuller		8 years ago
592	Pull together a few bits to start writing a Developer's Manu... Pull together a few bits to start writing a Developer's Manual. This pulls in the existing code style doc and Olaf's Principles Of Operation, and adds a set of links and a new doc describing how f.functions are implemented and dispatched. Hopefully it'll grow more information and better organization over time.	Matthew Fuller		8 years ago

Older »