summaryrefslogtreecommitdiff
path: root/xlib.c
Commit message (Collapse)AuthorAgeFilesLines
* major internal changes in representation of character cells.Rich Felker2006-10-291-2/+2
| | | | | | | | | | | | | we now use 12 bytes per cell instead of 10. however, this allows us to support 256-color mode (not yet implemented but the framework is in place) and to mix scripts when using combining characters. while the latter sounds ridiculous at first, being able to visibly see a combining letter attached to a [, ", or ' is extremely useful in scripting and regular expressions with some languages. some code is left slightly messy, but overall it's much cleaner now since struct uucell is now properly encapsulated.
* no need to check thisRich Felker2006-10-261-2/+0
|
* 10l: multiple-free!Rich Felker2006-10-261-1/+3
|
* basic, primitive paste support. it is blatently incorrect for non-utf8Rich Felker2006-10-261-5/+39
| | | | | | | locales, and will also fail when the selection owner cannot provide utf-8 encoded text. much more support code is needed, but at least this is a start.
* stupid omission; still a glitch though..?Rich Felker2006-10-251-0/+1
|
* major changes to input handling in preparation for pasting, which willRich Felker2006-10-251-26/+22
| | | | | | | | | | | | | | | | | | | | require support for large blocks of input: output which cannot be written to the tty (pty) is no longer discarded. instead, the display module is responsible for keeping the buffer alive until it has been entirely consumed. one simple (and smart) way to ensure this is to refuse to process events that could generate new input text as long as there is unwritten data. this works as long as there is an earlier layer of buffering (for fbcon, kernel; for x11, xlib). a direct-on-x-protocol display implementation will have to do its own buffer management. this architecture allows pasting to be performed direct from a buffer owned by the display module, which could reside in static or shared memory, or even as a memory-mapped file. it also facilitates reuse of the code in non-tty-connected settings.
* mistaken line left over from old codeRich Felker2006-10-161-1/+0
|
* don't know if this is a bug somewhere else, but for whatever reason iRich Felker2006-10-161-0/+1
| | | | | was getting non-null-terminated strings sometime?!
* remove server-side glyph rendering/cache entirely. client side is justRich Felker2006-10-151-66/+55
| | | | | | as fast or faster even with remote displays due to horribly inefficient X protocol...
* revert "performance increase" that actually hurt performance on betterRich Felker2006-10-151-30/+19
| | | | | | | X servers.. :( some better approach is really needed; at this rate, even client side images would be faster...
* large performance increaseRich Felker2006-10-151-19/+30
|
* preliminaries for input method supportRich Felker2006-10-151-2/+13
|
* better now, perhapsRich Felker2006-10-131-1/+4
|
* nonlatin input, attempt 2Rich Felker2006-10-131-2/+2
|
* try to support non-latin input.. does this work?Rich Felker2006-10-131-12/+26
|
* remove margin nonsense i got from reading bad sourceRich Felker2006-10-131-4/+3
|
* remove the horribly incorrect XImage handling and replace it withRich Felker2006-10-131-21/+16
| | | | | something much simpler that won't corrupt memory.
* x11 window resizing supportRich Felker2006-10-131-36/+52
|
* broken backspace; Xlib is stupidRich Felker2006-10-121-0/+1
|
* possibly fix crash on last glyph page?Rich Felker2006-10-121-2/+2
|
* missing prototypesRich Felker2006-10-121-0/+1
|
* 1% less messy/hackishRich Felker2006-10-121-6/+7
|
* first try at porting to X. very broken, minimal functionality,Rich Felker2006-10-121-0/+375
but it does work..