Commit Graph
761 Commits
Author SHA1 Message Date
RRFZ 90b447797a Treat only SQLite's ASCII whitespace as token breaks in CREATE TABLE parsing (R3.12) 2026-11-18 04:18:10 -05:00
RRFZ 58d5d35781 Recognize an escaped slash after a gitignore globstar (R3.12) 2026-11-18 04:09:29 -05:00
JG Tromp 788642e487 Offer Simple / Advanced on the Make screen, and always show Clean room / Deep there (R2.9) (#1286)
* Claim Simple / Advanced and Clean room / Deep on the Make screen (R2.9)

* Offer Simple / Advanced on the Make screen, and always show Clean room / Deep there (R2.9)

Simple mode's title bar leaves the mode switch out, so someone starting on
New project never saw that Advanced exists. The ways in now end with a
Simple / Advanced switch (on a line of its own when the window is too
narrow), using set_mode like Settings and the title bar; switching from
there stays on New project instead of opening the workbench, so the draft
stays in view.

The Clean room / Deep tiles appeared only once an original was picked. They
are now always in view: with nothing to recreate they are dimmed, can't be
picked, and say the choice applies once there is something to recreate.
Choice tiles report whether they are enabled to accessibility.

* Put back this PR's roadmap line, dropped by a merge of main, worded as review asked (R2.9)

A merge of main made in GitHub's conflict editor dropped the R2.9 line this
PR ticks. It's back, now saying Simple's title bar had no mode switch (only
Settings had one) rather than that Simple hid it everywhere but Settings.

* Describe #1286's two new tests in doc comments (R2.9)

* Measure the room Make's Simple / Advanced switch needs instead of assuming 240 points (R2.9)

Review of #1286: the row beside the ways in fitted the switch by a fixed 240-point width, set
for English. It's now measured from the switch and its name in the language shown
(widgets::segmented_width, from main's design system), so a longer translation moves the switch
under the ways in sooner. The narrow-window test also sweeps five window widths and checks the
switch and its name never overlap the ways in and stay inside the window.
2026-11-18 16:08:41 +07:00
RRFZ 66804ba597 Cover alias-expanded IntPtr, UIntPtr and dynamic controls in C# terminal fallback tests 2026-11-18 04:05:35 -05:00
RRFZ d04467530b Check exact tails before TypeScript fallbacks and skip external HTML references in js_app (R3.12) 2026-11-18 04:00:28 -05:00
BFlat f21b14d001 Merge pull request #2097 from Plastic-time/bugfix/r3.12-js-app-typescript-resolution
In a TypeScript project that hasn't been built, `import "./util.js"` means the file `util.ts`: the import names what the file will be called after building. Mapping such a folder left those imports unresolved, so the app's modules looked unconnected. An import or preload path ending in `.js`, `.mjs`, `.cjs` or `.jsx` now falls back to the TypeScript file beside it when no file of that exact name exists, and `.mts`/`.cts` files and more kinds of `index` file are found. The declaration files `.d.mts` and `.d.cts` are no longer taken for code, as `.d.ts` already wasn't.
2026-11-18 15:47:04 +07:00
BFlat e047eda07d Merge pull request #2095 from Plastic-time/bugfix/r3.12-js-await-sequence-import
Follow-up to #2083, which stopped an unawaited `import("x")` being taken for the module itself. Bundled code often writes the awaited form as `await (0, import("x"))`, and that shape was no longer recognised, so calls on the module it loads weren't recorded. An awaited comma sequence that ends in a dynamic import is the module again; the same sequence without `await` still isn't.
2026-11-18 15:27:48 +07:00
RRFZ bdf0854760 Fix js_ledger expectation source_node IDs, candidate deduplication, and same-owner binding aggregation (R3.12) 2026-11-18 03:24:45 -05:00
BFlat d60732c4c0 Merge pull request #2096 from Plastic-time/bugfix/r3.11-pseudo-indirect-call-target
A call through a register (`blx r3`, `call rcx`, `call [rdi + 16]`) reads that register to find its target. The pseudo-C treated the register's assignment as an argument being set up: it removed the statement and passed the function's address as a parameter, so `r3 = table[2]; (*r3)(1, 2)` came out as `(*r3)(1, 2, r2, table[2])`. An assignment the call itself reads is now kept as a statement and isn't folded into the arguments.
2026-11-18 15:19:46 +07:00
RRFZ 8213aecf60 Exclude .d.mts/.d.cts declarations and resolve TypeScript source specifiers in js_app (R3.12) 2026-11-18 03:08:03 -05:00
RRFZ 18b931b6b9 Report dedicated uninstall blocking diagnostics and UTF-8 remediations in setup (R2.2)
- Build dedicated `ActionKind::Invalid` blocking records in `plan::uninstall` when a client config cannot be read, fails syntax parsing, or refuses in-place entry removal instead of calling `inspect_client(c, &[])` and reporting `"Can't add our entry..."` / `"run setup again."`.
- Share `plan::read_text` (`ReadFail::InvalidUtf8` vs `ReadFail::Io`) across `plan`, `doctor`, and `uninstall` so non-UTF-8 configs advise saving as UTF-8 rather than checking file permissions.
- Add `uninstall_blocking_diagnostics` integration regressions and update `docs/ROADMAP.md` and `docs/STATUS.md`.
2026-11-18 03:04:07 -05:00
BFlat a955c188fb Merge pull request #2094 from Plastic-time/bugfix/r3.12-js-recover-trailing-dots-and-ext-case
Windows drops a dot at the end of a file or folder name, so `pkg./a.js` and `pkg/a.js`, or a host written `ex.test.` and `ex.test`, were planned as two files and written as one, and the second was refused. Dots at the end of each part of a path are now removed before names are compared, as dots at the start already were. A module id ending in an upper-case extension (`src/App.TSX`) also keeps it, where `.js` used to be added after it.
2026-11-18 14:58:22 +07:00
RRFZ 3238242472 Add await-sequence 23/24 depth-cap boundary regression tests 2026-11-18 02:52:30 -05:00
RRFZ 50212b66f0 Keep alias-expanded native integer args exact-match-only 2026-11-18 02:50:58 -05:00
RRFZ 631191861b Tokenize tight SQLite CREATE TABLE types, defaults, and table options (R3.12) 2026-11-18 02:48:41 -05:00
RRFZ cdfeff5aca Preserve indirect-call target and base registers during pseudo-C argument folding (R3.11)
- Remove the `Flow::Call | Flow::IndirectCall` exemption around `defs.remove(r)` and `consts.remove(r)` in `Lifter::block` so registers read by indirect calls (callee pointers, memory-indirect bases/indices, and AArch64 PAC modifiers) retain their assignments instead of folding into callee argument lists.
- Add four integration regressions in `crates/disasm/tests/pseudo_indirect_call_target.rs` and update `docs/ROADMAP.md` and `docs/STATUS.md`.
2026-11-18 02:43:56 -05:00
RRFZ d2078af0e2 Align sequence-await depth cap and expand regression cases 2026-11-18 02:40:25 -05:00
RRFZ 2d856f34a8 Preserve module identity for awaited sequence dynamic imports 2026-11-18 02:23:20 -05:00
BFlat 642179c6ab Merge pull request #2093 from Plastic-time/bugfix/r3.12-csharp-contextual-keywords-and-fallback
Follow-up to #2080. `nint`, `nuint` and `dynamic` are only keywords when nothing in scope has that name, but a conversion operator to a program's own type called `nint` or `dynamic` was read as the built-in and reported as missing; such a name is now looked up among the program's types first. An explicit `System.IntPtr` keeps meaning the built-in. And a conversion to `List<String>` where `String` is another namespace's type, imported without `using System;`, is matched to that type before `String` is tried as `string`.
2026-11-18 14:11:50 +07:00
BFlat bd0f0235c7 Merge pull request #2090 from Plastic-time/bugfix/r3.11-dataflow-mode-registers
Tracing where a value goes in a function leaves out the stack pointer and program counter, which every instruction touches. On 32-bit ARM those are `r13` and `r15`, which weren't on the list, so each `push`, `pop` and stack store added them to the graph as if they carried data. They are now left out for ARM and Thumb only; on x86-64, where `r13` and `r15` are ordinary registers, nothing changes. A 32-bit x86 call also no longer lists three register names that are never used there.
2026-11-18 13:58:47 +07:00
BFlat 51f1f2b34f Merge pull request #2092 from Plastic-time/bugfix/r3.12-dotnet-member-body-generic-context
Fields, properties and local variables of a generic .NET type already showed its type parameters by name, but an event's type and everything a method body refers to (calls, field accesses, the listed instructions) showed them as numbered placeholders: `System.Action<!0, !1>`, `Dictionary<!1, !!0>::Add(!0, !1)`, `box !1`. They are now read with the enclosing type's and method's parameters, and a call on a generic type shows the argument types of the type it is called on.
2026-11-18 13:58:46 +07:00
BFlat f1e70b10a2 Merge pull request #2088 from Plastic-time/bugfix/r1.5-gdb-mi-framing
A console command sent to GDB with a line break in it was written as two lines, so its second line ran as a command of its own; line breaks, tabs and other control characters are now escaped. A result that arrived late from an earlier command that had timed out could be taken as the current command's result or leave its error behind; only the result carrying the current command's number counts now. A stray `}` or `]` in a stop record no longer hides the fields after it, and `\a`, `\b`, `\f`, `\v` and `\xHH` in GDB's output are decoded.
2026-11-18 13:54:02 +07:00
BFlat f6c6543713 Merge pull request #2089 from Plastic-time/bugfix/r3.12-js-fallback-caps-and-truncation
When a script can't be parsed, a simpler reader lists what it can from the script's words. It listed the address of a `fetch("https://…")` call twice and repeated an address for every time it was written; it stopped noticing addresses at all after 4,000 other strings; and when its lists of strings, IPC uses or exposed names were full it dropped the rest without saying so. Each address is now listed once, addresses are found however many strings came first, and a list that was cut is named as cut.
2026-11-18 13:53:12 +07:00
RRFZ 5dc919fd74 Strip trailing dots from recovered module and web-script segments and match known module extensions case-insensitively (R3.12) 2026-11-18 01:46:44 -05:00
BFlat 3f8e1eb3de Merge pull request #2086 from Plastic-time/bugfix/r2.2-qoder-skills-and-windows-tilde
Someone who moves Qoder's configuration folder with `QODER_CONFIG_DIR` got its settings written there but the skill written under `~/.qoder`, where their Qoder no longer looks; both now go to the folder the variable names. On Windows, a path given in such a variable as `~\folder` was taken as a folder literally named `~` inside the home folder; it is now the home folder, as `~/folder` already was.
2026-11-18 13:43:33 +07:00
RRFZ c888defe2c Preserve explicit System.IntPtr/UIntPtr identity across contextual keyword reconciliation 2026-11-18 01:27:16 -05:00
RRFZ 9e7ffa3f47 Propagate generic context to CLI events and method-body tokens (R3.12) 2026-11-18 01:25:15 -05:00
RRFZ 8e84985f6f Resolve C# contextual type keywords and scope-first generic fallback (R3.12) 2026-11-18 01:19:13 -05:00
RRFZ 2a5af743a1 Filter ARM32/Thumb SP and PC in value flow and canonicalize x86 call clobbers (R3.11)
- Scope `dataflow::ignored(mode, r)` so `"r13"` and `"r15"` are filtered as SP/PC bookkeeping only in `Mode::A32` and `Mode::T32`, preserving `r14` scratch reuse on ARM and general-purpose `r13`/`r15` reaching definitions on `Mode::X64`.
- Trim the fallback 32-/16-bit x86 `clobbered` list in `dataflow::def_use` to canonical `&["rax", "rcx", "rdx"]`.
- Add `dataflow_mode_registers` regressions and update `docs/ROADMAP.md` and `docs/STATUS.md`.
2026-11-18 01:17:08 -05:00
RRFZ 6ac0c21454 Align JavaScript lexer-fallback list caps, URL deduplication, and truncation reporting with syntax-tree pass (R3.12) 2026-11-18 01:14:09 -05:00
BFlat d2f31c11ac Merge pull request #2087 from Plastic-time/bugfix/r3.12-gitignore-pattern-rules
Reading a source folder with its `.gitignore` files applied differed from git in four ways. A pattern starting with a backslash to mean a literal `*`, `[` or `?` was read as a wildcard. A name ending in an escaped space lost the space. `**` inside a name (`a**b.js`) crossed folders, where git treats it as one `*`. And a file under an ignored folder could be brought back by a later `!` rule, which git doesn't allow. Each now matches what `git check-ignore` reports.
2026-11-18 13:04:11 +07:00
RRFZ 2db47e2022 Keep colon-containing Windows tilde tails rooted under home 2026-11-18 01:02:07 -05:00
BFlat f0616948db Merge pull request #2085 from Plastic-time/bugfix/r3.12-dotnet-nested-generic-signatures
For a type nested in a generic type, .NET metadata lists the type arguments of the whole chain together. They were all written after the last name, so `Outer<int>.Inner` was shown as ``Outer`1/Inner<int>`` and `Outer<string>.Inner2<int>` as ``Outer`1/Inner2<string, int>``, and comparing C# source with the program reported such methods as different. Each part of the name now gets the arguments it declares. When the counts don't add up (obfuscated metadata), the arguments are kept together at the end as before.
2026-11-18 13:01:23 +07:00
BFlat 4a66500304 Merge pull request #2083 from Plastic-time/bugfix/r3.12-js-cooked-template-literals
A backtick string with no `${}` in it is a plain string, but several readers took its text as written in the source, escapes and all, or ignored it: ``require(`node\x3afs`)`` was recorded as that spelling, not `node:fs`, and a route, a preload setting or a bundle's module id written with backticks was missed. One helper now reads a quoted string or such a template by its value everywhere a string is expected. A template with an invalid escape, which is a syntax error unless tagged, is no longer reported as text.
2026-11-18 12:57:03 +07:00
BFlat 768d24a49e Merge pull request #2084 from Plastic-time/bugfix/r3.11-arm-pc-literal-refs
32-bit ARM code keeps constants and addresses in small pools beside the code and loads them with `ldr rX, [pc, #off]`. Those loads recorded no reference, so strings, data and imported functions reached that way had no cross-references, and the stubs that jump to an imported function (`ldr pc, [pc, #off]`) weren't recognised as imports. The address each such load reads is now recorded, the stubs are named after the function they jump to, and the pseudo-C shows these 32-bit loads as `uint32_t`, not `uint64_t`.
2026-11-18 12:46:33 +07:00
RRFZ 13947d37ec Follow QODER_CONFIG_DIR for Qoder skills and expand Windows ~\ paths in setup (R2.2) 2026-11-18 00:33:40 -05:00
BFlat e05b3b7242 Merge pull request #2082 from Plastic-time/bugfix/r3.13-share-path-collision-key
Recovering a bundle's modules and saving a capture's scripts each had their own copy of the code that decides whether two paths are one file on a disk that ignores case; both now use the one in the js crate. That key also treats `ß` and `SS` as the same, as a Mac disk does. And a file planned at a path that another file needs as a folder (`pkg.js` beside `pkg.js/child.js`) is renamed whichever of the two comes first, where the later one used to fail to write.
2026-11-18 12:21:45 +07:00
RRFZ 15eb535211 Format nested generic CLI signatures per segment arity (R3.12) 2026-11-18 00:18:42 -05:00
RRFZ a0ea1bc443 Honour gitignore backslash escapes, quoted trailing spaces, globstar segments, and excluded parent directories in Ignore (R3.12) 2026-11-18 00:09:09 -05:00
RRFZ 56e4c4478d Record ARM32 and Thumb PC-relative literal references and GOT jump slots (R3.11) 2026-11-18 00:06:33 -05:00
RRFZ e87104f45e Escape GDB/MI C-strings, clamp malformed tuple delimiters, and isolate command tokens (R1.5)
- Escape `\n`, `\r`, `\t`, and remaining control characters (as fixed three-digit octal byte escapes) in `quote`, and decode standard C escapes (`\a`, `\b`, `\f`, `\v`) plus `\xHH` hex byte escapes in `unquote` so console commands and stderr log lines stay single-line MI records and round-trip cleanly.
- Clamp unmatched closing braces and brackets at nesting depth zero in `fields` so malformed tuple delimiters outside quotes cannot hide subsequent top-level MI fields such as `reason`.
- Normalize CRLF line endings in `drain` and pass `expected_token` to `absorb` so late or untokenized result records from other commands cannot taint `c.result` or `c.error`, while clearing stale errors on non-error records in `take_output` and still recording `^exit`.
2026-11-18 00:02:12 -05:00
BFlat 00f426522c Merge pull request #2080 from Plastic-time/bugfix/r3.12-csharp-conversion-builtin-aliases
Follow-up to #2076, which stopped a program's own `App.String` being read as `string` but left two conversion operators unmatched against a .NET program's metadata: one to `System.Collections.Generic.List<String>`, and one to `String` in a type that also converts to an unrelated `Other.String`. A short type name in a conversion target is now looked up first in the enclosing types and namespaces among the types the program has; only when nothing there has that name is it read as the built-in (`String` as `string`).
2026-11-18 11:56:47 +07:00
BFlat 2a96867125 Merge pull request #2081 from Plastic-time/bugfix/r3.11-sbpf-ja32-and-jump-opcodes
In eBPF and Solana programs, `ja32` (opcode 0x06) is an unconditional jump whose distance is in the 32-bit immediate. It was decoded as a conditional branch using the 16-bit offset, so the jump went to the wrong place and the code after it looked reachable. It is now a plain jump to the right target. Three opcodes that no version defines (0x0d, 0x0e, and 0x9d before version 2) were read as `ja` or `exit`; they are now shown as unknown, while version 2's `return` (0x9d) is kept.
2026-11-18 11:56:40 +07:00
RRFZ 1904af648c Share full case-folded path and parent-directory collision keys across JS recovery and web script layout (R3.13) 2026-11-17 23:18:44 -05:00
RRFZ 6f6442e2d6 Decode sBPF ja32 (0x06) as a 32-bit jump and reject invalid BPF_X jump/exit opcodes (R3.11) 2026-11-17 23:17:01 -05:00
RRFZ bd8e224300 Track generic arity, predefined keywords, and resolved inner arguments in C# conversion lookup 2026-11-17 23:15:49 -05:00
BFlat 5e555c9c41 Merge pull request #2079 from Plastic-time/bugfix/r3.11-thumb-symbol-and-pointer-targets
A 32-bit ARM shared library has no entry point, so it was always decoded as ARM code, also when it was built as Thumb, and came out as nonsense. The mode is now taken from its function symbols, exports and function starts: Thumb when most of them carry the Thumb bit. In Thumb code that bit is also cleared wherever an address names code (functions that never return, exports, the entry point, code pointers and pointers found in data), so those match the functions they point at; on other processors an odd address is left as it is.
2026-11-18 11:15:34 +07:00
BFlat 50a9e7aa52 Merge pull request #2067 from Plastic-time/bugfix/r2.2-yaml-empty-and-flow-blocks
Setup refused a Hermes `config.yaml` whose server list was written `{ }`, `Null` or `NULL`, though each means an empty list; those are accepted now. A server entry written as a flow mapping over several lines can't be bounded by indentation, and editing one left its closing brace behind: setup and uninstall now stop and say to edit the file by hand rather than guess. An entry whose flow value is complete on its own line is edited as before, and its siblings are left byte for byte as they were.
2026-11-18 11:05:55 +07:00
BFlat c66f36f8db Merge pull request #2074 from Plastic-time/bugfix/r3.12-js-recover-case-collisions
Two modules recovered from a bundle whose paths differed only in letter case (`X.js` and `x.js`) were given both names. On Windows and on a default Mac disk those are one file, so writing the second was refused and the module was left out. Names are now compared without regard to case, also after a hash or a number has been added, so the second module gets a name of its own.
2026-11-18 10:53:09 +07:00
RRFZ 4fb286a00c Resolve unshadowed CLR built-in aliases in C# conversion targets (R3.12) 2026-11-17 22:43:25 -05:00