aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* fix some more off by one errorsHEADmasterJesse Luehrs2017-07-181-4/+4
|
* suppress some compiler warningsJesse Luehrs2017-07-171-2/+4
|
* bump libvt100Jesse Luehrs2017-07-171-0/+0
|
* fix off by one errorJesse Luehrs2017-07-151-2/+8
|
* don't call memcmp on nullJesse Luehrs2017-07-151-0/+5
|
* suppress fallthrough warningJesse Luehrs2017-07-021-0/+1
|
* move the mouse reporting response generation into libvt100Jesse Luehrs2017-06-022-60/+50
|
* implement motion mouse reportingJesse Luehrs2017-06-012-7/+34
|
* refactor mouse reportingJesse Luehrs2017-06-011-19/+25
|
* fix mouse reporting with lines in scrollbackJesse Luehrs2017-06-011-0/+1
|
* bump libvt100Jesse Luehrs2017-06-011-0/+0
|
* Revert "combine flushes when reading large amounts of data"Jesse Luehrs2017-02-124-61/+1
| | | | | | This reverts commit 6ee215a7797cbb15aed6136ccfcaf5daff5ae654. actually, the redraw_rate stuff should already be handling this
* combine flushes when reading large amounts of dataJesse Luehrs2017-02-124-1/+61
| | | | | on linux at least, you can only read 4096 bytes of data from a pty, so large writes would cause a lot of redraws
* another syntastic optionJesse Luehrs2017-02-121-0/+1
|
* avoid an allocation on every glyph string drawJesse Luehrs2017-02-123-10/+20
| | | | | probably not super meaningful compared to the cost of the cairo operations, but may as well
* optimize drawing sequences of ascii charactersJesse Luehrs2017-02-122-12/+117
| | | | bypasses pango for this common case, which should speed up drawing
* fix moving the mouse after making a selectionJesse Luehrs2017-02-112-6/+18
| | | | | previously it was always moving the existing selection around even if the mouse button wasn't held down
* only clear the selection if the contents changedJesse Luehrs2017-02-113-12/+18
|
* optimize drawing of runs of empty cellsJesse Luehrs2017-02-111-43/+64
|
* optimize runes_display_continue_string a bitJesse Luehrs2017-02-111-10/+5
| | | | | continuing is the common case, so this trades two compare+branches for two ORs and a NOT
* add makefile target for profilingJesse Luehrs2017-02-111-2/+9
|
* also handle ctrl+arrowsJesse Luehrs2017-01-271-0/+12
|
* remove the wide emoji hacksJesse Luehrs2017-01-265-6/+1
| | | | | they have been folded into modern versions of glib (>= 2.50.1) as part of unicode 9.0.0
* make wide emoji configurableJesse Luehrs2017-01-255-1/+7
|
* support input escapes for shift+special keysJesse Luehrs2017-01-211-4/+20
|
* handle double click to select strings of wide charactersJesse Luehrs2016-06-062-13/+30
|
* clean up the sample config fileJesse Luehrs2016-06-051-38/+51
|
* allow configuring double click rateJesse Luehrs2016-06-051-0/+3
|
* fix redrawing the screen after a visual bellJesse Luehrs2016-06-051-1/+1
|
* bump libvt100Jesse Luehrs2016-06-051-0/+0
|
* ignore profiling dataJesse Luehrs2016-06-051-0/+3
|
* rows, cols, not cols, rowsJesse Luehrs2016-06-051-1/+1
|
* allow storing reversed start/end for selectionsJesse Luehrs2016-06-051-7/+8
| | | | | | | only ensure they're in the right order when pulling the selection text out of it, but leave it reversed otherwise (or else selecting behind the initial starting point will actually move the starting point of the selection, which is wrong)
* avoid a race condition if the window closes while we're double clickingJesse Luehrs2016-06-051-0/+3
|
* no reason to flush the surface hereJesse Luehrs2016-06-051-1/+0
| | | | since we're immediately calling runes_window_flush anyway
* no reason to go through an event hereJesse Luehrs2016-06-055-60/+41
| | | | | | | | i was only doing this before because libuv was using threads, and so i needed a way to trigger redraws on the main thread. if there are other reasons that we might want to delay redraws (to avoid blocking the event loop, for instance), we should just do that by scheduling a 0-timeout timer or something.
* this is unnecessary now that we have rate limitingJesse Luehrs2016-06-051-29/+0
|
* simplifyJesse Luehrs2016-06-051-4/+2
|
* don't acquire the selection until we have something selectedJesse Luehrs2016-06-052-38/+54
| | | | | | clicking in a terminal window shouldn't be sufficient to remove an existing selection from another window (although it should clear the selection in its own window)
* this is probably unnecessaryJesse Luehrs2016-06-051-1/+0
|
* allow more easily running a development version alongside a real oneJesse Luehrs2016-06-054-8/+16
| | | | | | | this just changes the window class (so that i can make the development version use floating windows while not screwing up the layout for my real terminals) and the runesd socket location (so that i can test runesd at all while i'm already running a real runesd).
* have the display own the entire selectionJesse Luehrs2016-06-054-42/+48
| | | | rather than just the selection location
* manually send selection clear events when starting a new selectionJesse Luehrs2016-06-051-0/+12
| | | | | | | this happens automatically if we're taking over the selection from external windows, but doesn't happen if we're just taking the selection over from another window that we share a display connection with (like other runesc windows)
* reset the timer here before doing things that take timeJesse Luehrs2016-06-051-8/+8
|
* this should also count as being handledJesse Luehrs2016-06-051-0/+1
|
* update selection on right clickJesse Luehrs2016-06-051-0/+3
|
* implement double click to select words and triple click to select linesJesse Luehrs2016-06-054-4/+173
|
* allow canceling timers before they expireJesse Luehrs2016-06-052-2/+21
|
* bump libvt100Jesse Luehrs2016-06-041-0/+0
|
* bump libvt100Jesse Luehrs2016-06-041-0/+0
|