Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
50 most recent check-ins
|
2024-04-25
| ||
| 08:13 | (build) suppress warnings, wrong version of tui_lua committed previously Leaf check-in: 81cad85759 user: void tags: trunk, master | |
| 04:54 | (tui) add support for .title to readline popup hints check-in: 2cdc5564e9 user: void tags: trunk, master | |
|
2024-04-23
| ||
| 04:07 | (build) add fossil-settings ignore to skip external/ and build/ check-in: e1879486a5 user: void tags: trunk, master | |
|
2024-04-22
| ||
| 03:26 | (tui) readline: better position management for suggestion area check-in: ade7fd5e6b user: void tags: trunk, master | |
|
2024-04-21
| ||
| 16:57 | (lua) define ENTER/EXIT for all ALT calls when USE_TRACY is set We still don't get proper frame tracking for some reason which still blocks the utility of it all, requires a better debug setup than what I have access to now. check-in: 7935aed6cb user: void tags: trunk, master | |
| 14:22 | (vrbridge) swap nreal implementation for OHMD PR The reversing effort done in Monado is more thorough than our quick bringup, so better to swap through that. Since Monado's driver interface is a bit more involved, there turns out to be a OHMD PR that won't go anywhere on behalf of the project being stale, but we can go with that and some later massage to also handle Xreal Air 2 / Pro variants. check-in: 8767451e66 user: void tags: trunk, master | |
| 14:20 | (build/core) pull in shmif-tui for future text_surface call This adds stub types and the API entry point, but lacks core still. The point is to get a simpler path to experimenting with server-side text while at the same time prepare the ALT to A12 path to retain TPACK format in buffers for remote GPU work. Ideally this should be able to eventually replace most render_text calls as used in Durden, to later facilitate a tpack-tui-window which works with a shared atlases with optional MSDF variant. check-in: dbfc1496aa user: void tags: trunk, master | |
| 07:36 | (tui/lua) allow glob on directory to resolve to opendir also extends the fstatus table with omitted ctime/atime/mtime check-in: bdb9bc7128 user: void tags: trunk, master | |
|
2024-04-19
| ||
| 06:25 | (tui-lua) add fchmod, fchown and extend fstat This is still driven by discoveries while implementing features in cat9, the extended fstat requires a third bool argument for detail=yes as it imposes several more calls and returns a populated table which is unwanted for many cases. check-in: 0b10b59c6d user: void tags: trunk, master | |
|
2024-04-15
| ||
| 02:26 | (tui) re-introduce copy window This existed before the move to server side text processing. The form it is in now is that every tui window is forced a 'COPY_WINDOW' input label which takes and resolves the active front buffer into tpack and tunpacks into a new window check-in: f1d78cff0a user: void tags: trunk, master | |
| 02:25 | (core) minor toggle change for default tracy check-in: f9bd1e6589 user: void tags: trunk, master | |
|
2024-04-12
| ||
| 17:45 | (core) split out the tracy parts to a separate unit and compile in cpp This lets us pull in the tracyLua.hpp for better insight and export into the appl layer. check-in: 5e45783fd4 user: void tags: trunk, master | |
| 15:07 | (shmif) re-type from enum to int in arcan_shmif_signal for cpp check-in: 1604c42d11 user: void tags: trunk, master | |
|
2024-04-11
| ||
| 10:53 | (agp) regression in synchronous readback last modification missed the edge condition of specific readback into vidp check-in: c879ef299e user: void tags: trunk, master | |
| 04:01 | (shmif/tui) add fd to local path resolver and map to on_bchunk in tui-lua This is a necessary evil to avoid extraneous bgcopy actions when handling an incoming descriptor. On supported OSes it tries to match a descriptor to a local file path and matching inode. This makes it possible for the recipient to forego creating a local copy first in order to pass to a consumer which requires access to a filepath. check-in: 892094c89a user: void tags: trunk, master | |
| 00:28 | (tui) fifo flags regression and suppress some 'unused' check-in: 5595751071 user: void tags: trunk, master | |
| 00:18 | (core/tui) synch all vendored form of nbio to match between lash and core check-in: d637c353ac user: void tags: trunk, master | |
|
2024-04-10
| ||
| 11:52 | (tui) bchunkstate request should provide extension for wildcard check-in: aeba0a5e01 user: void tags: trunk, master | |
| 11:51 | (shmif) changes to internal fbo backing store rotation and signalling check-in: 3be79bbb2c user: void tags: trunk, master | |
| 11:50 | (platform/lua) add glob_resource overloaded form that supports asynch globbing check-in: 8ea8400dac user: void tags: trunk, master | |
|
2024-04-05
| ||
| 21:56 | (build) vendor fts because the glibc/musl won't ever go anywhere check-in: b089f95018 user: void tags: trunk, master | |
| 21:29 | (core) transform-mask action invalid on first created object, reported-by: vimpostor check-in: edc02f398f user: void tags: trunk, master | |
| 21:28 | (doc) typo breaking manpage generation, reported-by: vimpostor check-in: fa5cf0ffba user: void tags: trunk, master | |
| 21:17 | (shmif) adjust headers and type definitons to play better with cpp check-in: cd3a20ccc9 user: void tags: trunk, master | |
|
2024-04-03
| ||
| 20:38 | (tui/lua) change how HINT-flag is terminated and avoid off-by-one on HINT-fl but no field check-in: 831bc5bd4c user: void tags: trunk, master | |
| 18:27 | (shmif/tui) extend _border drawing with a blend flag, and add F1 toggleable readline help for suggestion hint-sets. check-in: d82ba5d2bb user: void tags: trunk, master | |
| 18:26 | (core) dirty state propagation test This is a non-fix but useful regardless as we get slightly better tracing info. The full issue is still that if an inner rendertarget has an attachment that is dependent on a vstore in an outer rendertarget, and that vstore gets updated as part of a frameserver poll sequence pre-render of the outer rendertarget, and there is nothing going on in the inner rendertarget it will not update until some other event inadvertently triggers it. This gets more obvious at low refreshrates. check-in: cdf5423114 user: void tags: trunk, master | |
| 17:28 | (doc) fix typo check-in: 9fc33060f1 user: void tags: trunk, master | |
|
2024-03-28
| ||
| 09:38 | (core) avfeedframe ordering edge condition This is related to [6c2a968ead] but not entirely. The problem was that a client could race the STEPFRAME socket signal with the atomic flag and end up clearing a frame and consuming the event for the next one at roughly the same time. If the eventloop in the client then solely relied on a pollset with no other events coming (basically triggered by the single-consumer thread setup Xarcan uses for output segment to native window mapping) then it would stop processing. check-in: 3d74d99e29 user: void tags: trunk, master | |
| 08:31 | (lua) fix [6c2a968ead] by reworking rectarget setup check-in: 959a6cb669 user: void tags: trunk, master | |
|
2024-03-26
| ||
| 21:23 | (lua) rectarget window size regression due to platform/... no longer probing store on w=0,h=0 check-in: d712885171 user: void tags: trunk, master | |
|
2024-03-24
| ||
| 13:34 | (platform/egl-dri) previous fix omitted a line, try again check-in: f9412b9ba2 user: void tags: trunk, master | |
|
2024-03-23
| ||
| 20:39 | (platform/dri)(credit:bitcompost) prevent kernel-framebuffer container leak check-in: 1ebd87ac02 user: void tags: trunk, master | |
|
2024-03-21
| ||
| 20:23 | (shmif) synch CURSORHINT description suggested labels to better match w3c check-in: def2a11b06 user: void tags: trunk, master | |
|
2024-03-20
| ||
| 20:26 | (egl-dri) mark current set display mode as primary check-in: 6beb4a947c user: void tags: trunk, master | |
|
2024-03-18
| ||
| 14:25 | (lua) refactor open_nonblock for easier separation / namespace management check-in: 2d63b481dd user: void tags: trunk, master | |
|
2024-03-13
| ||
| 00:03 | (meta) merge offline branch work check-in: 5e834308da user: void tags: trunk, master | |
|
2024-03-12
| ||
| 14:49 | (doc) target_seek clarification check-in: beee765f58 user: void tags: trunk, master | |
|
2024-03-11
| ||
| 23:23 | (tui/lua) expose cursor_io output tagging to lua bindings check-in: 864c1a8087 user: void tags: trunk, master | |
| 23:21 | (tui) add entrypoints for cursor-io announcement check-in: 8a74ee2f8b user: void tags: trunk, master | |
| 23:09 | (shmif/lua) extend BCHUNKSTATE event with flag for cursor (DND) tag check-in: c696eef43b user: void tags: trunk, master | |
| 07:02 | (shmif/decode) add option to block-for-activate shmif cont acquired with NOREGISTER and patch into decode, this fixes a regression in decode-pdf. check-in: eb98490187 user: void tags: trunk, master | |
|
2024-03-09
| ||
| 16:16 | (shmif) wrong change committed, early-out condition inversed. check-in: 2dff9cdf4b user: void tags: trunk, master | |
| 14:08 | (egl-dri) plug drmPropertyBlob for atomic pageflip leak check-in: ff999cacb4 user: void tags: trunk, master | |
|
2024-03-08
| ||
| 20:21 | (shmif) check for missing context in shmifext_setup check-in: 479a6795e7 user: void tags: trunk, master | |
| 16:23 | (shmif) set init.rate on outputhint in preroll check-in: 77bac5b41b user: void tags: trunk, master | |
|
2024-03-07
| ||
| 09:01 | (lua) open_nonblock() used wrong usermask preventing opening _APPL (only _APPL_TEMP) check-in: c26b9a214d user: void tags: trunk, master | |
|
2024-03-01
| ||
| 10:26 | (net) prevent .log filename truncation on wide-pids from misuse of %.nd precision check-in: 1560234752 user: void tags: trunk, master | |
|
2024-02-24
| ||
| 14:26 | (tui) check for existing input_key handler on mouse-wheel event check-in: b55678b661 user: void tags: trunk, master | |
| 09:53 | (doc) update README.md with fossil information check-in: a4a2b18fe6 user: void tags: trunk, master | |