aboutsummaryrefslogtreecommitdiffstats
Commit message (Expand)AuthorAgeFilesLines
* split: guard toggle behind prompt-active; clear _split in the pilot resetblasty2 days2-0/+3
* split-view phase 1: side-by-side listing <-> pseudocode layoutblasty2 days3-3/+179
* decomp: fix stuck 'decompiling' spinner when re-opening a loaded functionblasty2 days2-0/+14
* decomp: prettier 'decompiling' overlay — animated braille spinnerblasty2 days1-2/+31
* decomp: restore the 'decompiling…' overlay on F5/Tab from the listingblasty3 days2-7/+27
* palette: fix results width (were compounded narrower than the modal)blasty3 days1-2/+3
* palette: give the command palette side padding (80% width, max 120)blasty3 days1-0/+3
* palette: replace the stock Ctrl+P command palette with real ida-tui actionsblasty3 days2-0/+102
* ux: omit the splash logo when the terminal is too small for itblasty3 days1-1/+7
* hex: freeze the cursor's screen position on scroll (not clamp-to-edge)blasty3 days2-26/+38
* ux: render the logo.ans ANSI-art splash in the loading overlayblasty3 days2-2/+61
* hex: cursor follows the scroll (stays visible on wheel/scrollbar)blasty3 days2-4/+25
* hex: click to place the byte cursor (hex + ascii panes)blasty3 days2-0/+55
* docs: refresh README for the worker-only architectureblasty3 days1-16/+27
* xrefs: show fine-grained kind (call/jump/read/write/offset) in the dialogblasty3 days4-8/+93
* listing: shift+home jumps to the start of the instruction textblasty3 days1-0/+24
* listing: <home> jumps the cursor to start-of-line (pairs with <end>)blasty3 days1-2/+3
* listing: <end> jumps the cursor to end-of-line, not goto-bottomblasty3 days1-1/+2
* docs: sweep for the worker-only reality (drop mcp supervisor/spawn.sh/--db)blasty3 days4-53/+30
* mcp: delete the ida-pro-mcp transport, supervisor, and mcp-only testsblasty3 days21-2236/+67
* mcp: collapse app + launcher to worker-onlyblasty3 days3-330/+54
* refactor: extract shared error hierarchy + Session into idatui/errors.pyblasty3 days6-69/+99
* tests: pass url=""/db=None to IdaTui in the worker pilot pathblasty3 days1-1/+1
* tests: add --worker <binary> to run the pilot suite against the idalib workerblasty3 days1-5/+16
* worker: inject custom tools on startup (self-sufficient, no spawn.sh needed)blasty3 days1-0/+20
* deprecate: make the idalib worker the default backend; mark mcp path for removalblasty3 days6-11/+47
* worker: wrap non-dict tool returns as {"result": ...} to match MCP shapesblasty3 days1-1/+5
* worker: spawn under the IDA python (has ida_pro_mcp), not the TUI'sblasty3 days1-4/+33
* worker: surface the real startup failure (not just "code 1")blasty3 days2-7/+41
* app: --backend {mcp,worker} — run the TUI on our idalib worker (migration s...blasty3 days2-66/+115
* worker: idalib worker + WorkerClient (drop-in for IDAClient) — migration st...blasty3 days3-0/+383
* experiments: add a unix-socket idalib worker as the 3rd bench columnblasty3 days1-29/+187
* experiments: in-process idalib vs mcp-transport spike (throwaway)blasty3 days2-0/+348
* robust: survive a lost server connection instead of crashing; auto-reconnectblasty3 days1-2/+95
* nav: baseline horizontal scroll at 0 when jumping to a target columnblasty3 days1-4/+5
* ux: fix truncated loading-splash messageblasty3 days1-2/+1
* ux: block with a "finding xrefs…" overlay while gathering xrefsblasty3 days1-6/+71
* nav: snap an xref-into-pseudocode jump to the line that holds the symbolblasty3 days1-5/+40
* nav: xref jump into pseudocode lands on the referenced symbol, not column 0blasty3 days1-1/+5
* nav: land the decompiler cursor on the function name, not column 0blasty3 days1-5/+11
* ux: label the startup wait as auto-analysis, not "starting a server"blasty3 days1-3/+8
* ux: start the analysis server from inside the TUI (no more dead window)blasty3 days2-50/+46
* launch: sweep a crashed worker's stale DB locks before starting the supervisorblasty3 days1-0/+9
* ux: show TUI chrome + a "loading…" overlay during binary open/analysisblasty3 days2-16/+104
* fix: keep the pseudocode cursor/scroll when a rename re-decompilesblasty3 days1-1/+12
* nav: land the cursor on the ref's column, not column 0blasty3 days1-6/+26
* fix: refresh-after-edit keeps the edited line on screen (rename now visible)blasty3 days1-2/+13
* fix: Tab decomp->listing reuses the nav entry so edits stay putblasty3 days1-5/+25
* fix: Tab back from a decomp jump returns to the listing cleanlyblasty3 days1-2/+11
* nav: jumping from the decompiler stays in the decompilerblasty3 days1-8/+61