aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* seed: regenerate after merging demo-fixes (image c079adab)Gravatar zinc 12 days2-5/+5
|
* merge work/demo-fixes: analyzer ownership fixes (back links, loop-built ↵Gravatar zinc 12 days38-31/+435
|\ | | | | | | lists, .u inference), Io.print any args, block-end borrow messages
| * prescan: guess an owned return only from a local declared from new/heap, not ↵Gravatar zinc 12 days3-16/+28
| | | | | | | | a global assigned one (fixes lsp self-check); xref baseline for the new writers
| * lsp test: block-end borrow message and related infoGravatar zinc 12 days1-0/+9
| |
| * tests: alias_scope expects the block-end message; plain_return now covers a ↵Gravatar zinc 12 days3-3/+12
| | | | | | | | forwarding owned return called early; direct owned return before its body is accepted
| * analyzer: a .u (declared or inferred) function can't return a borrowed ↵Gravatar zinc 12 days6-1/+45
| | | | | | | | parameter or a value derived from one; fixtures
| * std.Io.print/println take any mix of printable arguments (compiler expands ↵Gravatar zinc 12 days6-7/+63
| | | | | | | | per type); docs, conformance
| * TODO: implicit namespace at the topGravatar zinc 12 days1-0/+2
| |
| * analyzer: a back link (container's owner stored into it) stays a plain ↵Gravatar zinc 12 days10-1/+72
| | | | | | | | reference, fixing a compile hang on a<->b and a leak; ending an owner kills its owned children too (fixes use after reassigning a container); escape fixtures
| * analyzer: infer .u from returns of fresh objects in the body (prescan guess, ↵Gravatar zinc 12 days4-1/+66
| | | | | | | | then checked as .u)
| * analyzer: an outer owner moved into a deeper container can take it back ↵Gravatar zinc 12 days9-4/+140
| | | | | | | | (loop-built lists); block-end frees name the owner (why 9) in messages and LSP related info
* | merge work/actors: wasm seed bootstrap (seed/MANIFEST), shared copy-on-write ↵Gravatar zinc 12 days52-89/+1266
|\ \ | |/ |/| | | actor memory, top-level hold blocks on all backends and lsp
| * bench: rss probe path fixGravatar zinc 12 days1-1/+1
| |
| * bench: actor peak-rss probe (N=0/10/100 on zy, zyi, zym)Gravatar zinc 12 days4-0/+46
| |
| * seed: regenerate for the hold sources (image d792bd4f)Gravatar zinc 12 days2-5/+5
| |
| * hold: copy hold.h with micro.c everywhere; xref baseline for hold globals; ↵Gravatar zinc 12 days13-19/+37
| | | | | | | | lsp keyword completion yields to a same-named symbol; compiler message avoids the word actor
| * docs: hold blocks in LEARN.md (36-hold), spec 5.16, ACTOR.md; TODO: m:n ↵Gravatar zinc 12 days6-1/+97
| | | | | | | | scheduler and follow-ups
| * lsp: hold and shared hovers, completion of the keywords, shared variables ↵Gravatar zinc 12 days4-2/+67
| | | | | | | | marked in hover and completion; tmLanguage keywords; hold_lsp_test.js
| * test.sh: hold fixtures on interp, micro, x86-64 jit and llvmGravatar zinc 12 days1-0/+12
| |
| * hold: fixtures (accept, reject, race, rank order) and tests/drivers/hold_test.pyGravatar zinc 12 days13-0/+141
| |
| * hold blocks: top-level hold { shared let ...; fn ... } monitors; shared ↵Gravatar zinc 12 days6-3/+208
| | | | | | | | values kept by the host and copied in/out at the outermost call, reentrant per-block locks in declaration order (core, micro)
| * actors: micro workers map a copy-on-write snapshot too; inherited input is ↵Gravatar zinc 12 days3-15/+113
| | | | | | | | only the unread bytes; worker stacks 4 MB
| * actors: remaining tls tables and the input buffer allocated on use; a worker ↵Gravatar zinc 12 days2-19/+47
| | | | | | | | inherits only unread input
| * actors: large per-thread compiler tables mapped on first use instead of ↵Gravatar zinc 12 days1-18/+75
| | | | | | | | static tls (13 MB of every worker's stack mapping)
| * actors: worker memory is copy-on-write against a snapshot of the parent ↵Gravatar zinc 12 days1-10/+100
| | | | | | | | (memfd/shm), o3 entry tables handed over sparse
| * onboard: the wasm seed bootstrapGravatar zinc 12 days1-6/+12
| |
| * boot: build the compiler from a pinned wasm seed (seed/zc.wasm, WASI) under ↵Gravatar zinc 12 days9-6/+321
|/ | | | wasmtime or node, proven by the fixpoint; seed/MANIFEST hashes; make seed; legacy c bootstrap kept as make genome-c
* merge work/borrow: borrow returns before definition, borrows clause, borrow ↵Gravatar zinc 12 days32-44/+514
|\ | | | | | | messages and LSP diagnostics, mkgenome header mismatch values
| * xref: baseline records the borrow globals (end_why, end_whom, arg_site, ↵Gravatar zinc 12 days1-3/+13
| | | | | | | | local_shadow, xshadow, effect_shadows, dgrel)
| * mkgenome: on a header mismatch name FTOP/ZFP, print the image's and the ↵Gravatar zinc 12 days2-7/+28
| | | | | | | | map's values, and point at a stale src/runtime/rt.zc or map
| * lsp: borrow errors ranged on the use, with related information for where the ↵Gravatar zinc 12 days9-6/+154
| | | | | | | | owner ended and was declared; vscode client passes relatedInformation through
| * analyzer: check borrow returns of forward-called functions at the unit's ↵Gravatar zinc 12 days16-10/+212
| | | | | | | | end; borrows clause for prototypes, definitions, externs and function types, checked against the body
| * analyzer: use-after-end messages name the borrower, the owner and how the ↵Gravatar zinc 12 days12-21/+110
|/ | | | owner ended (freed, moved, .d parameter, push, global, reassignment), with fixtures
* merge work/stdio: std.Io and one-level module completionGravatar zinc 12 days11-10/+184
|\
| * lsp: prefer the rt import fix when a library group wraps the same nameGravatar zinc 12 days1-1/+2
| |
| * lsp: complete a module path one level at a time (std. lists groups, ↵Gravatar zinc 12 days2-5/+36
| | | | | | | | std.File. and std.Io. their members)
| * std: add std.Io (print, println, nl, put_*, put_f, get_c, eof, read, ↵Gravatar zinc 12 days10-4/+146
|/ | | | getline) with test, learn example and docs
* xref: baseline records the prover's writers of ft_ln (bin, kw_for), xub and ↵Gravatar zinc 12 days1-0/+3
| | | | xubat (bin)
* todo: overnight freelist-inline done-gate met with trees timingsGravatar zinc 12 days1-1/+3
| | | | | | Mark native freelist pop/push inline done (c31a07d/63ff1aa), record x86-64 jit trees ~3× win and llvm/C comparison ratios; clear the overnight still-open line.
* x86-64 jit: inline new_k, recycle_leaf and own fast paths, in o3 regions and ↵Gravatar zinc 12 days6-7/+306
| | | | | | | | | | | | | | plain blocks a call to rt.new_k / rt.recycle_leaf with a constant class, or to rt.own, does its class-list pop or push (or owner store) inline and goes on after the call; an empty list, memdebug, a held range, a block outside memory, a translated word or a type/class mismatch takes the ordinary call. in a region this replaces the o3_invoke round trip. o3 frees of leaf objects now use recycle_leaf too (new_k stamps the type inside o3 since b86db60). ZYNOFASTALLOC=1 turns it off; ZYPERFMAP=1 writes a perf map of translated code. tests/programs/o3_freelist.zc: reuse, zeroing, an empty list, untyped neighbours, memdebug; jit vs ZYNOFASTALLOC vs ZYO3INTERP.
* rt: recycle, new_k and own without helper calls or || chainsGravatar zinc 12 days1-18/+58
| | | | | | | | | | perf (ZYPERFMAP) on plain trees new/free showed recycle, new_k and own taking about 70% of the time, mostly in calls to isobj, paylen, bdrop, bclass and fpush and in the truth values plain code builds for && and ||. recycle writes isobj, paylen and the class-list push out, calls scan_ptr only for 8 or more words left, and own and new_k test with nested ifs. frame_control grows to 22 words: a tag, &dbg and the entries of new_k, recycle_leaf and own for the x86-64 jit's inline fast paths (next commit).
* todo: overnight 5:27 queue verified; find_replace hardening notedGravatar zinc 13 days1-1/+4
| | | | | drop-check trees/sieve timings vs e67671d, reassignment and o3 drop/GC checks, plain new_k confirmation, and 1db57f7; freelist-inline still open.
* find_replace: cover o3 function prologues on every backendGravatar zinc 13 days7-7/+99
| | | | | | | | arm64, llvm and the region-interpreter rdrop path now watch the two prologue words the way the x86-64 jit already did, so icon.replace_function's lit+jmp patch retires a compiled callee (and clears fn_nat) instead of leaving a stale stub. three new icon regressions: repeated o3 replace, failed publication keeps the old entry, and a direct call after warming the region.
* todo: overnight prover/rcall/alloc statusGravatar zinc 13 days1-1/+1
|
* prover: do not restore length facts after a call-loopGravatar zinc 13 days3-17/+19
| | | | | | | Restoring undirtied lafacts when the body might free was wrong: push can grow an array without dirtying its name, so a[99] after push warned against a stale length 0. Keep the old lafact_keep rule; reorder explain_prover so the constant/affine cases run before the touch() loops.
* prover: len/cross-array/fixed-new/affine proofs; keep facts past call loopsGravatar zinc 13 days5-19/+44
| | | | | | | | Rebased the prover WIP onto main. lafact_keep restores undirtied allocation facts even when the loop body might free (so a later touch() no longer drops every length fact). proven_p uses ptr_still so an outer call keeps the check. entry_fact still supplies the size for the "might free" explain reason. bounds_prover.zc and explain_prover.zc land with test.sh coverage.
* wip: prover gaps (inline len bound, cross-array, rt.new depth, fixed new, ↵Gravatar zinc 13 days7-43/+172
| | | | affine i*n+k); unbuilt and untested
* o3: use new_k (size-class free lists) like plain codeGravatar zinc 13 days2-4/+7
| | | | | | | Constant-size typed new inside o3 now calls rt.new_k instead of always taking new_t/new_raw. new_k and recycle_leaf skip the o3 argument-fatten path the way free/recycle already do. trees o3 new/free on the x86-64 jit dropped from about 4.1 ms to about 2.7 ms per iter (matches plain).
* x86-64 o3: direct o3-to-o3 calls (rcall_x64 / rfinish_x64)Gravatar zinc 13 days2-16/+184
| | | | | | | | | Port arm64's rcall/rfinish to the x86-64 region jit: call the callee's entry stub through fn_nat with a built frame, return via linkstk, and fall back to o3_invoke when the stub is missing or the frame/depth check fails. Clear fn_nat on rforget and cover the function prologue so icon.replace_function drops the region (o3_caller_replace). vm_until saves jit_depth around nested bytecode, matching arm64.
* todo: destructors closed with what landed; overnight handoff notes (triage ↵Gravatar zinc 13 days1-1/+5
| | | | verdict, file and tsc installed, genome 0x70000)
generated by cgit v1.2.3 (git 2.39.1) at 2026-10-27 14:52:04 +0000