| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Clear mailbox of unmatched processes | Naveen Michaud-Agrawal | 2023-11-23 | 1 | -0/+2 |
| | | |||||
| * | display: WIP: Refactor into files; use wishes, not Display:: | Omar Rizwan | 2023-10-06 | 1 | -11/+13 |
| | | | | | | | | | Expose shader stuff to user programs for the first time. Mostly works, but display thread and main thread run much slower (28fps display, 70fps main). Also introduce dict getwithdefault (helpful for rest options). | ||||
| * | process: Allow running multiple On process blocks on process | Omar Rizwan | 2023-10-06 | 1 | -6/+18 |
| | | | | | | | | | | Turn off share wishes/claims default for subprocesses, so you (for now) have to explicitly specify sharing if you want any to happen. On process now only spawns process if it doesn't exist but always runs the code block (which now happens through a statement instead of directly on init). | ||||
| * | WIP: Rewrite peering to use shm instead of websockets | Omar Rizwan | 2023-08-15 | 1 | -5/+8 |
| | | | | | Huge performance increases, but crashy. | ||||
| * | WIP: process/peer reworking | Omar Rizwan | 2023-07-21 | 1 | -32/+38 |
| | | | | | | | | | | 1. Use a permanent blacklist so processes from unmatched matches can't leak statements if they connect after the destructor is gone 2. Have subprocesses die on disconnect from the main process TODO: need to handle races between multiple camera processes. | ||||
| * | test/process: Add test for disposal of briefly-appearing process | Omar Rizwan | 2023-07-21 | 1 | -0/+3 |
| | | |||||
| * | Clean up process unmatch handler | Omar Rizwan | 2023-06-28 | 1 | -17/+3 |
| | | |||||
| * | Clean up unmatch and sharing so it actually only runs on subproc | Omar Rizwan | 2023-06-19 | 1 | -18/+21 |
| | | |||||
| * | Use m/s: format for stringifying ids. Make subproc share all | Omar Rizwan | 2023-06-18 | 1 | -3/+1 |
| | | |||||
| * | Introduce fork-zygote process system. | Omar Rizwan | 2023-06-18 | 1 | -48/+76 |
| | | | | | | | | | | | | | | | Also refactors/separates ::nodename to be ::thisNode (this computer) and ::thisProcess (this OS process). All tests should pass. Zygote-based processes are much faster to start up since they don't need to recompile C code or reload Folk. They still need to connect to Folk via WebSocket for now, and the automatic management we'd want (bidirectional, based on both match lifetime and process lifetime) isn't working yet. Some of the node vs. process distinction in naming still needs to be cleaned up, too, I think. | ||||
| * | Fix handleReadable | Omar Rizwan | 2023-06-02 | 1 | -0/+1 |
| | | |||||
| * | Merge branch 'main' into osnr/camera-pipeline | Omar Rizwan | 2023-06-02 | 1 | -1/+1 |
| |\ | |||||
| | * | All tests pass | Omar Rizwan | 2023-06-01 | 1 | -3/+4 |
| | | | |||||
| | * | Use one-off namespaces so inner scope can't stomp outer scope | Omar Rizwan | 2023-04-26 | 1 | -1/+1 |
| | | | | | | | | | | | and so that we can keep blowing away namespaces instead of doing cleanup | ||||
| * | | Use chan open for process execution | Charles Chamberlain | 2023-06-01 | 1 | -10/+14 |
| | | | |||||
| * | | Getting one camera frame from a thread | Charles Chamberlain | 2023-06-01 | 1 | -1/+11 |
| | | | |||||
| * | | Fix bug in process naming: choose unique names in same file | Charles Chamberlain | 2023-06-01 | 1 | -0/+4 |
| | | | |||||
| * | | Implement bidirectional peering & receive statements | Omar Rizwan | 2023-05-25 | 1 | -2/+4 |
| | | | |||||
| * | | Allow bare When/Claim in subprocess | Omar Rizwan | 2023-05-25 | 1 | -1/+4 |
| |/ | |||||
| * | Guard against variable stomp when you call Step inside a Step | Omar Rizwan | 2023-03-18 | 1 | -1/+1 |
| | | |||||
| * | Add support for `shares all claims`; fix Commit in subprocess | Omar Rizwan | 2023-03-17 | 1 | -1/+1 |
| | | |||||
| * | Use a closed variable instead of global in the commit | Omar Rizwan | 2023-03-17 | 1 | -0/+1 |
| | | |||||
| * | Introduce new process test (doesn't work yet) | Omar Rizwan | 2023-03-13 | 1 | -17/+0 |
| | | | | | | (Prep for putting camera in a managed process, then doing incremental camera) | ||||
| * | Basic process statement sharing test passes. | Omar Rizwan | 2023-02-14 | 1 | -0/+20 |
| | | |||||
| * | More cleanup/refactoring of peering. | Omar Rizwan | 2023-02-14 | 1 | -4/+6 |
| | | | | | Start trying to test IPC statements. | ||||
| * | Refactor peering. Refactor process. | Omar Rizwan | 2023-02-14 | 1 | -8/+11 |
| | | |||||
| * | process: Refactor to use per-process namespace | Omar Rizwan | 2023-02-14 | 1 | -32/+43 |
| | | |||||
| * | process: Show stdout/stderr as label. | Omar Rizwan | 2023-02-14 | 1 | -10/+19 |
| | | |||||
| * | Refactor sharing to have a more general 'peers' system. | Omar Rizwan | 2023-02-11 | 1 | -1/+1 |
| | | | | | | | | | | | | | | | | | | This is in prep for interprocess communication. For now, it is just used to implement pretty much the same generational laptop-sharing-to-table scheme, although done with explicit share-patterns instead of implicitly based on assertion, which is nice. Also make `Claim` and `Wish` store the calling $this instead of `someone`. Also start on `Do` for controlling side effects. Also move back to the old way of collecting matches & build it into main -- this is used as a hack for negation (so we can explicitly retract the Pi root statements when a laptop connects) | ||||
| * | Move process logic into lib/process.tcl | Omar Rizwan | 2023-02-10 | 1 | -0/+30 |
