A pin's menu says "Open folder", as the project page does, for the same action. A game's chosen folder is shown like every other chosen folder: shortened, cut to fit its row and whole on hover, so a long path no longer widens the page.
A server that took the request but didn't start answering in time now says so instead of "Couldn't reach". When the AI runs on this computer (localhost, 127.x or ::1), the advice says a large model may still be loading or reading a long prompt, and a server that can't be reached is told to check it's running here, with nothing about the internet connection. Remote servers keep their advice.
`LC_MAIN` gives a Mach-O executable's entry point as a file offset. It was mapped through the last segment starting at or below that offset, so a `__DATA` holding only zero-fill sections, which has no file bytes but records file offset 0 like `__TEXT`, took the entry point and placed it outside the code. The offset now maps through the segment whose file bytes contain it, and an offset no segment holds gives no entry point.
A SquashFS metadata table cut inside a block's header or payload now says so with a warning and marks the recovery incomplete, instead of stopping silently. Only blocks that were read are mapped, so the blocks before the cut stay readable and nothing points into the missing one.
A default filled into an older row is now read the way SQLite reads it into that column: a TEXT column reads a number as text, NUMERIC and INTEGER columns read text that is a number as that number, a REAL column then makes it a REAL, and BLOB, untyped and STRICT ANY columns keep the literal as written. Numbers too large for a REAL stay unsupported instead of showing as infinity.
Ollama fits a prompt that is too long for its context by dropping the oldest messages; once the person's request is among them, a Qwen 3.8 model refuses what's left with 500 "no user query found in messages". That refusal is now reported as a context overflow, saying how to give the model a larger context or send less, and it is no longer retried, since the same request would be refused again. Other server errors are still retried.
Fixes#1245
Source map lookups find the section and the segments at a generated position by binary search over the sorted lists instead of scanning them all, keeping every mapping that shares the matched column.
Source map VLQ fields now decode the whole signed 32-bit range, including the reserved encoding of the minimum, reject values beyond it, and accept redundant zero groups without overflowing a shift.
Some linkers write a version string with wType 0 (binary), and the PE reader dropped it, such as a DLL's FileDescription. A binary leaf at the string depth is now decoded as UTF-16LE text, while a binary node that holds a child node is still walked as a container.
Rows stored before ALTER TABLE ADD COLUMN now get the added column's default when it is a hex integer (`DEFAULT 0x10`, kept as SQLite's 64-bit two's-complement value) or a single literal in parentheses (`DEFAULT (5)`, `DEFAULT ('hi')`), instead of NULL with an "unsupported DEFAULT" note. Expressions and nested parentheses still show the note.
A column's declared type keeps the comma inside its parentheses, so `DECIMAL(10, 2)` is listed as written instead of `DECIMAL(10 2)`, in the parsed columns and the schema overview; only the comma that separates columns is dropped.
Schema rows were decoded with the 4 KiB cap used for every text value, so a CREATE TABLE longer than that lost the columns after the cut, its rowid alias and its WITHOUT ROWID marker. The schema is now read whole for parsing (up to 1 MiB per statement and 32 MiB per walk, after which text is cut as before), while the statement shown is still cut at 4 KiB on a character boundary.
A build carrying on without a new request deleted the API builder's kept conversation before starting, so it could never resume and always studied again. The conversation is now kept whenever the run is continuing, and only a genuinely fresh build (no continuation and no finished study) starts a new one; a conversation from the other mode is still refused when it loads.
SQLite stores a REAL with no fractional part as an integer and turns it back into a REAL when it reads a column with REAL affinity. Rows now read the same way, including WITHOUT ROWID keys and filled-in defaults of added columns, so `1.0` no longer shows as `1`. Affinity follows SQLite's order, so a `FLOATING POINT` column, which contains INT, stays an integer column.
Solana program relocations are read at the table's declared entry size, which must be at least the record's own size, and each record must be complete; a short entry size or a truncated table is reported instead of fields being read from the next record.
Before a certificate's fields are used to verify a signature, the algorithm OID must be the first element of its sequence and valid, the key and signature bit strings must be non-empty with no unused bits, and the issuer and subject must be sequences. A certificate that fails these checks is not taken apart for verification instead of yielding an empty algorithm or misread bytes.
Adds tar tests for PAX records with a signed length, an empty key or an equals sign inside a path value: malformed framing stops the listing, and `path=a=b` names the member `a=b`.
A cpio member's name must end in a NUL (extra NUL padding is still allowed), and an archive counts as complete only when its trailer is empty and fully present. An unterminated or invalid name, or a trailer with data or cut short, stops the listing with a warning instead of the archive being reported complete.
The XML plist reader now takes exactly one value, inside a closed `<plist>` when there is one, with only whitespace, comments and declarations around it: unclosed roots, extra values, stray text, non-empty `<true>`/`<false>` and a `<key>` outside a dictionary are errors instead of being read past. A DOCTYPE's internal subset is skipped as one declaration, without fetching or expanding entities. Invalid UTF-8 and a half UTF-16 code unit are reported instead of being replaced.
An index source map's sections now need a valid offset and either an inline map or a `url`, in ascending order; malformed section tables are an error instead of defaulting to offset 0. A section that points to an external map, which is never fetched, keeps its start, so lookups past it no longer borrow mappings from the inline section before it.
The check that a module has one code body per declared function now counts every declaration in the function section, not only the ones the inventory kept, so a module with more functions than the listing limit no longer gets a false "function declarations but bodies" warning.
A packed bundle member whose stored or unpacked size differs from the size its manifest records is now an error instead of being returned as if it were complete.
Source map and external debug URLs in WebAssembly custom sections are read as WebAssembly names: invalid UTF-8 is rejected instead of being replaced, and long inline source map URLs are kept whole.
A function's local count now includes every local declaration group, not just the first 10,000; the inventory limit still applies to the number of functions listed.
An API builder's kept conversation now records whether it ran in Deep mode or Clean room (`transcript.json` version 2), and is only resumed in that same mode; a version-1 transcript counts as Deep. A run without a new request whose transcript came from the other mode studies again instead of carrying on, and a chosen extractor is no longer started in Clean room.
The Android binary XML reader stays inside the document's declared size, reads a node's fields after the header length it declares, and requires each attribute record and typed value to fit within its own element. Nodes that are too short or attribute tables with an impossible layout mark the document truncated with a warning, instead of borrowing bytes from a sibling or from past the document's end.
Keyed archives now follow object references inside arrays, inline dictionaries and objects that are only an alias for another, both in the object list and when projecting an object. A `$top` that isn't a dictionary is no longer taken for an archive, and the text a projection copies from repeated references shares one 64 MiB budget.
The SQLite reader found a table's column list at the first `(` in its CREATE TABLE, so a table named `"export (1)"` or `[Sheet1 (2)]` lost its columns, rowid alias and WITHOUT ROWID order. The list now opens at the first parenthesis outside quotes.
A relocation's symbol name is read only from a linked symbol table and string table of the right types, and only when the symbol index falls inside that table, so relocations keep their record but no longer borrow names from neighbouring sections.
Solana program section headers are read from records of the declared entry size, which must be at least 64 bytes, and a truncated section header table is reported instead of fields being read from the next record or past the end.
A contract's CBOR metadata trailer is split off only when it decodes completely. Text that isn't UTF-8, or arrays and maps longer than the reader keeps, now invalidate the trailer, so a plausible length suffix alone no longer removes instructions from the analyzed bytecode.
A GNU build ID is read only when its description, after the padded owner name, fits inside the note region, so it can no longer borrow bytes from whatever follows the region.
When a section or segment holds more file bytes than it maps into memory, the padding past its virtual size no longer gets a virtual address, and reads at an address stop at the mapped end instead of running into the padding.
Pointer reads follow the image's width: 16-bit binaries read two-byte pointers instead of four, and an unknown width reads nothing instead of guessing.
A GNU base-256 number with its sign bit set is negative, which an unsigned tar header field such as a member's size can't be; it is now rejected instead of being read as a huge size.
Resource compilers write an empty version string with no value at all, and the PE reader dropped those. Strings such as `OleSelfRegister`, whose presence is what matters, are now kept with an empty value.
A doubled quote inside a quoted column name now stands for one quote instead of ending the name, and single-quoted names, which SQLite accepts, lose their quotes. Keys that name such columns find them, so rowid aliases and WITHOUT ROWID order hold.
The safetensors reader checks tensor sizes for the 4-, 6- and 8-bit float types too, refuses a header over 100 MB, warns about metadata values that aren't strings instead of showing them as JSON, shortens long architecture and name fields, and reports tensor data ranges that overlap, leave gaps or leave bytes unclaimed at the end.
The binary plist reader accepts the `bplist0x` versions and reports others as unsupported. An unknown simple marker, a date marker of the wrong width or a dictionary key that isn't a string is now an error, instead of being shown as null or as an invented key.
A WITHOUT ROWID row stored before ALTER TABLE ADD COLUMN has fewer fields than the table has columns. Those rows now show the added columns' DEFAULT values, as rowid tables already did, instead of NULL.
A comma or quote inside a comment in a CREATE TABLE no longer splits or swallows column definitions, a comment before or right after a column name isn't read as part of it, and a comment inside a definition stays out of its type, so INTEGER PRIMARY KEY still marks the rowid alias.
docs/opencode.md now says that runs need OpenCode 2.x, that the install script and npm's `opencode-ai` still give 1.x, and that 2.x comes from npm's `@opencode/cli`, the desktop app's bundled CLI or the GitHub releases, with how to point the backend at it. It also warns not to start OpenCode during a build, since its shared service takes over the build's session. STATUS.md records the Linux live runs, and the roadmap gains the two problems they found.
JFFS2 extents and nodes, ROMFS payloads and header pointers, cramfs names, SquashFS metadata tables and file blocks, and tar PAX records now stay inside the sizes their images record. Out-of-bounds or truncated parts are reported as warnings and mark the recovery incomplete instead of being read past or silently dropped. PAX records need their terminator and a bounded name, and split octal fields in tar headers are rejected.
On Windows, a .cmd wrapper whose only command is one quoted, absolute .exe path followed by %* is started as that .exe, so its arguments skip cmd.exe (no % expansion, no 8191-character cap, line breaks allowed). Any other wrapper still runs as it is. The test that starts the .exe is Windows-only and isn't run by the Linux gate.
Fixes#1237
An API backend's timeout_secs now bounds the wait for its answer to start as well as the whole answer, so a request queued behind another tool's on a shared Ollama isn't cut off at five minutes; unset keeps five minutes to start and an hour in all, and a wait that runs out says to set it.
Fixes#1239
In Settings → AI, a card for an AI found on this computer lays its buttons out first and wraps the name and reason in the width they leave, so a long reason no longer runs under the Add button.
OpenCode model discovery, which also runs before every OpenCode turn, now uses --standalone unless the person chose a --server. Plain opencode api started OpenCode's shared service, which resumes every unfinished session on the computer, in any project, and carries on with its commands.
Part of #1115.
A VIRTUAL generated column stores no field in a row, so stored values no longer shift one column left: such a column is shown as not read, with a note, in rowid and WITHOUT ROWID tables alike, and a rowid alias after it is filled again.
The older, cut-off copy of the standalone-server sentence in opencode.md's configuration section goes; the newer copy, which also says where the file goes over SSH, stays.
Five STATUS rows that merges had kept twice, with different updates in each copy, are now one row each, the most complete copy (OpenCode models and thinking now says once that SSH runs get the tools), and opencode.md loses its older doubled paragraph. A pitfall says how to merge a conflicting table row.