Aider's adapter refuses a Clean room run in validate_access, but only the ACP path called it, so rebuild --clean-room with the Aider backend still started Aider and its unchecked reads. run_cli now checks the run's access mode before the setup check, so an agent that can't hold a Clean room run is refused on the native path too, before its program starts. Only Aider and Grok (ACP-only) override the check, so no other agent changes.
Since #1718 a run reads the project from disk when it starts, but the queue still decided whether a waiting run must build alone from the copy it was queued with. A Builder the person picked while it waited (an API backend, which must work alone) was then started beside the runs still working. The queue now reads the project when a run comes off it and hands that same copy to the run.
one_turn's only callers are Unix-only tests, so clippy with -D warnings failed on Windows with "function one_turn is never used"; CI runs on Ubuntu and didn't see it. The helper now has the same #[cfg(unix)] as its callers.
A run that waited in the queue, or one the rebuild tool started as the project's last run was ending, worked from a copy of the project taken before that run's last saves; its worker read the session from that copy and could save it back over the newer one, undoing the session the earlier run recorded. A run now reads the project from disk as it starts, keeping the given copy only if that fails.
A map with malformed mappings or section offsets was reported unreadable, although its sources list is usually whole, and the list is all an application graph keeps. A new sourcemap::parse_sources reads the sources, content, names and ignore list without decoding mappings or checking section offsets; the application graph re-reads a map parse refuses for those reasons this way, lists its sources with a source_map_mappings_unreadable limitation, and compares them as usual. A map whose decodable mappings name a source past the end of its list, or with a malformed section, is still unreadable.
A source map that was malformed, had an undecodable data URL or was listed but not loaded left its sources looking absent, so a source comparison reported them as confirmed missing. Such a map is now marked unreadable with a limitation, and the comparison reports otherwise unmatched sources as uncertain with partial coverage. A stand-alone .map beside an analysed script is read like a referenced one; other .map files stay listed and unread, so they can't make a comparison uncertain or match code that doesn't ship.
* feat(updates): add manual release checks and verified native updates
* chore: bump version to 0.6.1
* fix(flatpak): honor the launcher version choices
* fix(updates): report restart outcomes without UI process side effects
* fix(i18n): translate updater controls in the complete catalogs
* test: document native updater startup as UI plumbing
Follow-up to #1703. A sections array of non-objects could cost several times the old parser's memory, skipped values weren't checked as a full read checks them so some broken maps read as valid, entries a repeated key replaces still used the budget, and the analyzer read a map refused at a limit as one with no sources. Index-map sections now count against the entry budget and non-map sections stay small, skipped values get the same string, number and nesting checks, a replaced list is freed and gives its entries back, the default limit is 4,000,000 so a million sources with content fit, and a map refused at any of the parser's limits leaves its sources uncertain with a limitation, while a malformed map is read as before.
#1001 brought back the old open copy of the packaged Deep-build checklist line next to the ticked one and cut the ticked one off after "from the"; the cut-off line is removed and the full line ticked again, since docs/releasing.md has the check. The open line about the macOS remote-query test is removed too: its fix (#949) is already ticked.
The source-map parser read the whole JSON into a value and then copied every sources, sourcesContent and names entry, so a 64 MiB map of empty strings built about 21 million strings twice before anything was checked. It now reads only the fields it uses, straight from the JSON, and counts every sources, sourcesContent, names and ignore-list entry against one limit (2,000,000 by default) shared across sections; a map with more is refused with TooManyEntries as soon as the limit is passed, without reading the rest.
A project placed by a relative path gave engines and programs relative paths, but the child already runs inside app/, so Unity's -projectPath, Godot's --path, Unreal's -project and -archivedirectory, option runs (--screenshot, --self-test), console run checks and Play pointed inside app/ a second time. They now resolve the project path first; a program named by a bare PATH command is left as it is.
The source-map parser bounds the inventory it builds, but source recovery's manifest repeated every expanded source path in its outputs, so a long shared sourceRoot could still retain tens of megabytes of paths. Recovery now charges each original's source path and output path to the inventory's entry and byte budget, shared across inputs; once it's reached, further originals are left out, and the manifest, result, summary, a limitation and partial coverage say so. A pitfalls entry records the rule.
Craft Anything runs OpenCode 2 (opencode run --standalone). OpenCode 1.x has no --standalone, so the flag check saw it was missing, said "trying it anyway", and the build sat on OpenCode's usage text (#1111, #1693). When OpenCode's help was read and lacks the flags, the turn now stops at once saying the opencode found looks like 1.x, with how to install OpenCode 2 or choose it in Settings -> AI. A help that couldn't be read still lets the run go on, and a refused program is checked again next turn, so an updated OpenCode is seen. Other agents keep "trying it anyway".
oid_text, which names certificate fields and policies in decoded code requirements, showed malformed OID content as a different, valid-looking OID: an unterminated last subidentifier was dropped, a subidentifier padded with a leading 0x80 octet was accepted, empty content was an empty string, and more than 64 arcs were cut short. These now read "invalid", and a first subidentifier longer than eight octets is split into its two arcs like a short one.
Four open lines in the roadmap repeated a nearby line that is already ticked (the folder a build command ran in, what usage.json left out of a local build, a flaky helper test, and OpenCode's MCP server over SSH), so cargo xtask pending listed work the code already does. They are removed; each outcome stays in its ticked line.
What a builder gets from each coding agent (our tool server, our approval policy, resume, notes during a turn, a message longer than the command line) was spread over each adapter's arguments, environment and ACP session. cli_agents::grants now states it per adapter, takes resume and long messages from the adapter itself, and tests hold every row to what the adapter actually hands a builder; an adapter without a row fails them, and unchecked cells say NotYet with the gap in words.
When a source map's inventory was cut by its limits, the source-to-bundle comparison searched only the retained names and reported an omitted source as confirmed unmatched. Sources with no observed match now go in a separate uncertain_sources list with the reason source_map_inventory_truncated, the result says the inventory was truncated, and the evidence carries a limitation and partial coverage; positive matches are unchanged.
Distinct source identities such as src//main.js and src/main.js normalize to the same recovery path, so recovery kept the first and dropped the rest; a file and a folder of the same name, and separate inputs with the same name, lost sources the same way. Ordinary paths are unchanged; a source whose path would collide now goes to its own sources-collisions/<input>/<source>/ folder, and the manifest records each one's input and source-map index. A pitfalls entry records the rule.
A small source map could expand to gigabytes when its sourceRoot was copied into every source name, and the full list was built before the application analyzer kept 5,000 of them. Source lists are now built within an entry and byte budget shared across sections and an application's maps, checking each expanded path's length before allocating it; when only a prefix fits, the result says it was truncated and the analysis notes it as a limitation. A pitfalls entry records the rule.
The working card ("Working on it" in a build's feed, "Working on your answer" in the chat) drew its elapsed clock with a right-to-left layout after the words had wrapped across the whole row, so on a narrow card the clock covered the end of the detail line. The card now measures the clock and keeps its width out of the words' column, so the words wrap before it; a UI test checks they don't overlap.
Word likeness ignores one-character words, so an output made only of numbers had no words on either side and bare 4.5 against 9.0 passed as identical. When both outputs are only numbers and there is no second observation of the original, the check now compares the numbers in order and says which differ; output with words beside its numbers is compared as before.
When a program printed more than its output cap, its error said "what it printed last" from the first bytes kept and never said the rest was dropped. The process runner now also keeps the last 16 KiB of a stream once it passes its cap; failure messages, ssh diagnosis and build checks quote that end and say the output was cut, while tools that parse output still get its start.
After #1682, a run on a reattached project still used the builder picked on Make instead of the project's own, replaced the live conversation of a project already building, and cleared the Make draft even when the run didn't start. It now uses the saved builder (asking for one if it's gone), shows the running build's conversation and says it's already building, and clears the draft only once the request is running or queued. The run remains the only writer of the project's feed.
Since #987, picking a folder that already holds a project attaches to it, but the create path then overwrote its saved builder, asset budget, generated-assets choice and stack, replaced its conversation, and dropped the typed request. An attached project now keeps everything the person set and its saved conversation, and the request is added as its follow-up, in the app and the CLI; the CLI refuses --build-with, --generate-assets and --existing for such a folder with a fix to use instead.
The console run check compared words after dropping one-character tokens, so bare results like 4.5 and 9.0 looked empty and 12 matched -12. Study now runs the original twice per input; the check requires the numbers that were the same in both runs to match in order (keeping signs, decimals, exponents, radix and large integers), ignores numbers that varied between them, such as timings and PIDs, and says so in its note. Without a second observation it compares words only. Program names are replaced only as whole words, so they can't eat into numbers.
The roadmap asked for #1030's parent/descendant and startup-timeout cleanup tests to be run on Windows. They passed five runs in a row on Windows 11, and no processes they started were left running afterwards; the line is ticked with what was checked.
Source lookups and recovered source lists dropped sourceRoot when a source started with / or contained ://, giving the wrong original source. A non-empty root is now prefixed to every source before it is read as a URL, as ECMA-426 describes; an empty root still means no root, since generators commonly write "" for that.
Files reaching the third level of classic block pointers (above about 64 MiB with 1 KiB blocks) were left unrecovered. The reader now follows the triple-indirect pointer, keeps the full logical span of sparse subtrees at every level, and still refuses files whose mapping blocks are unavailable.
Method fingerprints wrote float constants as JSON numbers, which can't hold NaN or infinity, so methods differing only in such constants got the same fingerprint. Nonfinite constants are now written as their IEEE-754 bits, and single-precision NaNs keep their original operand bits, which widening to double could change.
SquashFS reads one XZ stream per block, but the block decoder refused a complete first stream whenever bytes followed its footer, so the file's data was lost. It now checks the stream the decoder consumed as a complete XZ stream on its own and returns it, ignoring what follows, and still never returns output from a stream that failed to decode.
The sBPF inspector capped symbols and relocations at 400,000 records, but reaching the cap only ended the current table, so each further SYMTAB/DYNSYM or REL/RELA table could add another record past it. It now stops visiting further tables of that kind once the cap is reached, keeping the earlier records and their order.
A cramfs file whose block-pointer table lay outside the available filesystem bytes was left without contents, but nothing said so, and the image still counted as completely recovered. Such a file now gives a warning and marks the image incomplete, keeping the file listed.
#1660 put notes an earlier run left back in the waiting queue, but a run with no message didn't reset the counts of unchanged turns and of checks that found no working build, so Continue on a build stopped by either limit stopped again before its builder read the notes. Taking such notes now resets both counts first, saving the check count on the project as it is on disk so a queued run doesn't undo what the run ahead of it saved; if they can't be saved, that's logged and the run starts anyway.
Target sniffing decides which tools are suggested. Two cases it already gets right had no test: a gzip stream named .txt stays Unknown, because the bytes decide the kind and not the name, and an empty .txt stays Unknown, because there is nothing to read.
Big-endian JFFS2 images were detected and then refused. The reader now takes the byte order from the first node's magic, after any erased space, and reads node, inode and directory integers in it, so the same tree and contents come out in either order; a node in the other byte order stops the scan with a warning.
newarray accepted any element-type byte and multianewarray accepted zero dimensions, both of which the JVM refuses. The decoder now stops at a newarray whose type isn't one of the eight primitive array types or a multianewarray with zero dimensions, as for other invalid instructions.
Any object that was the string "$null" was read as the archive's nil marker, so a real string with that text disappeared. Only the archive's first object, where the marker lives, is now read as nil; a "$null" string anywhere else stays a string.
DNS labels showed dots, backslashes and unprintable bytes as they came, so different wire names could look the same; they're now escaped (\. \\ \DDD). HTTP is reported only for a complete start line: a request with a known method, a target without control characters and an HTTP/1.x version, or a response with an HTTP/1.x version, a three-digit status and an optional reason phrase. The packet tools' warning about unreadable parts now says what may be missing.
In a ZIP64 archive with data in front of it, the locator's stored offset can miss the end record, and the classic fallback then counted the ZIP64 record and locator as prefix bytes and lost the central directory. Before falling back, the reader now looks for the record within 64 KiB before the locator, and takes it only when it is single-volume, agrees with the classic record and has a central directory right before it; a warning says it was recovered.
oid_text shifted each arc left by seven bits with checked_shl, which only fails for shift amounts of 64 or more, so an arc too big for 64 bits silently lost its high bits and showed as a different number. Arcs now build with checked multiplication and addition and stay at the largest value once they overflow.
Small ext4 inline files had their contents read as block addresses, so they failed to recover or came out as zeros, and a filesystem with only the inline-data feature was labelled ext2. Regular inline files of up to 60 bytes are now read from the inode, within the usual byte budget; longer inline files and inline directories are still reported as not recovered.
After a bad extent header or an unreadable child block, the extent walker returned a partly filled map, and recovery showed the missing ranges as zeros. Extent nodes now have to be complete and consistent (magic, entry counts, child depths, non-empty extents), within a shared work budget, or the file is left unrecovered with a warning.
LZ4 SquashFS images were refused although the project has a bounded raw LZ4 decoder. SquashFS metadata, data blocks and tail fragments now decode with it, treating the size as an upper limit, while Unity bundles keep requiring the exact decoded size.
Data blocks were concatenated whatever their length and the result trimmed at the end, so a short or corrupt first block could shift the bytes after it while the file looked complete. Each raw or decoded block, sparse blocks and the last partial block included, must now cover exactly its part of the file, or the file is left unrecovered with a warning.
The application graph is built from the assembly's inventory, members and native boundaries but only passed on the members' warnings, so a warning only the inventory or the native mappings raised (such as trailing bytes in an attribute's signature) disappeared from it. The graph now combines all three, keeps each view's first warning in view, and removes repeats before its 20-warning limit.
Directory and file pointers in a version-two cramfs image could reach bytes appended after the filesystem, so unrelated trailing data could fill in recovered contents. The size recorded in a version-two header now bounds every read, in either byte order; an image shorter than that size is read as far as it goes and marked incomplete, and version-one images are read as before.
The tokenizer matched punctuators against a four-byte string window, which came out empty when it ended inside a multi-byte character, so a===日本 split the operator. It now matches on the bytes from that point, so ===, >>, => and !== stay whole before any character.
Any ext inode that wasn't a directory or symlink was read as a regular file, so device, socket and FIFO fields could be taken as block pointers. The inode's mode now decides its kind: special and unknown inodes are listed as other entries without reading their fields as data, and a directory entry's type hint can't override the inode.
The certificate reader took any [0] field inside SignedData, kept non-X.509 certificate choices, and applied the certificate limit per field. It now reads the certificates field only in its declared place after the version, digest algorithms and content, returns only X.509 (SEQUENCE) choices, and applies the limit once.
#1644 sent an earlier run's waiting notes as the new run's message and deleted their files, so a run that stopped during setup lost them, and a note file that couldn't be read was dropped silently. Such notes now become the new run's own waiting notes, kept on disk until the builder takes them or handed to the next run like any note; a notes folder that can't be read is moved aside instead of deleted.