Skip to content

chore(deps): bump the rust-deps group across 1 directory with 6 updates - #25

Open
dependabot[bot] wants to merge 1 commit into
mainfrom
dependabot/cargo/rust-deps-ca0d415744
Open

dependabot[bot] wants to merge 1 commit into
mainfrom
dependabot/cargo/rust-deps-ca0d415744

Conversation

@dependabot

@dependabot dependabot Bot commented on behalf of github Nov 16, 2026

Copy link
Copy Markdown
Contributor

Bumps the rust-deps group with 6 updates in the / directory:

Package From To
serde_json 1.0.151 1.0.154
office_oxide 0.1.16 0.1.17
syn 3.0.6 3.0.10
thiserror 2.0.20 2.0.21
zip 9.1.3 9.5.1
proptest 1.11.0 1.12.0

Updates serde_json from 1.0.151 to 1.0.154

Release notes

Sourced from serde_json's releases.

v1.0.154

v1.0.153

v1.0.152

Commits
  • 20572a6 Release 1.0.154
  • a91bfaf Fix RandomState links in map documentation
  • 3c932c0 Merge pull request #1333 from razmser/no-std-preserve-order
  • 4b2d726 Release 1.0.153
  • 83335ee Touch up PR 1345
  • 4b4810a Resolve range_plus_one pedantic clippy lint
  • 8af71cc Merge pull request #1345 from 00200200/fix-invalid-utf8-error-column
  • 452afa2 Correct multiline raw UTF-8 error positions and simplify indexing
  • c9f1e3b Release 1.0.152
  • 854fc70 Touch up PR 1347
  • Additional commits viewable in compare view

Updates office_oxide from 0.1.16 to 0.1.17

Release notes

Sourced from office_oxide's releases.

v0.1.17 | Every finding of the 2026-09-27 audit — 19 issues, 165 checklist items — plus ten resource-bound defects reported privately, fixed test-first: each with a minimal synthetic document built in code, a test that fails before the fix and passes after. A two-arm sweep of 10,215 unique real Office documents against v0.1.12 then found nine regressions introduced by those fixes; all nine are fixed below, each with its own regression test, and the re-run sweep shows no crash, no timeout, no unintended new failure and no real slowdown.

Added

  • Pictures can be held as their stored streams, not inflated at open (lazy_pictures, opt-in). Opening a .docx, .pptx or .xlsx inflated every picture and kept it for the life of the document, although text, Markdown and HTML without images never read them: a 22 MB .docx of 39 page images held 341 MiB after open. With the new ParseOptions::lazy_pictures (or read_options::set_lazy_pictures, --limit lazy_pictures=1, or the option by name in every binding), a deflated picture that inflated to at least twice its stored size is kept as that stream and inflated only when it is used. On that file plain_text() peaks at 84 MiB instead of 404, the IR, to_html() and the IR JSON at about 395 instead of 696, and Markdown with images at 873 instead of 1,175, with every output byte for byte the same. Each picture is still read and checked at open, under the same limits and with the same reports, and only one is inflated at a time. The IR still copies each picture into every image that draws it, in both modes; sharing them, which needs a new field on ir::Image, waits for 0.2.0. New API: core::picture::PictureBytes (a shared, cheaply cloned handle with len, bytes and ptr_eq), Image::bytes(), Image::byte_len(), Image::has_bytes(), DocxDocument::picture / pictures, PptxDocument::slide_picture and XlsxDocument::picture. Off by default; without it nothing changes.

  • Large output without a host-language string, in every binding. A Python str is four bytes per character once one character is outside the Basic Multilingual Plane, a .NET string and most JS strings two, so the 450 MB IR JSON of a picture-heavy .docx cost a Python process 2.8 GiB. New forms hand over UTF-8 bytes or stream to a file: in Python plain_text_bytes(), to_markdown_bytes(), to_markdown_with_images_bytes(), to_html_bytes(), to_ir_json_bytes() and write_text / write_markdown(images=...) / write_html / write_ir_json, each to a path or a binary file object (1.5 GiB and 0.7 GiB on that file; what a file object's write raises is raised as itself, and a short write is completed); in Node doc.toBuffer(kind) and doc.writeTo(path, kind); in Go doc.Bytes(kind) and doc.WriteFile(path, kind); in C# doc.ToUtf8Bytes(kind), doc.WriteTo(Stream, kind) and doc.WriteTo(path, kind); in WASM doc.toBuffer(kind) (a Uint8Array); in C office_document_render_bytes and office_document_write_to_path. The output is byte-identical to the string forms encoded as UTF-8; the string methods are unchanged.

  • Reading time is checked against input size, in the tests and in the fuzzer. The limits bound sizes and counts, not time, and nothing caught the next algorithm that grows faster than its input. A new test suite (tests/test_time_scaling.rs) reads each shape that was once quadratic (a wide docx table, unclosed Markdown openers, a replacement across many runs, many clashing sheet names, many sheets, a large .xls, a long .doc) at two sizes through the public API and fails when four times the input costs ten times the time (a quadratic path costs sixteen). The fuzz jobs now fail an input that takes more than 10 s (-timeout=10; libFuzzer's default was 20 minutes). A cooperative step budget was prototyped in the worksheet event loop: its cost was within the noise of the timing, but a counter on parse events bounds only work linear in the input, which the size limits already bound, so it was not added as a limit. The README states that there is no time limit and that untrusted input should be read under a process timeout.

  • What a limit does, before a document is read: Limits::derived(), office-oxide limits and the options objects. The library derives eleven bounds from three of the eight limits by fixed fractions (picture and padding budgets, the write estimate, the share of package_bytes embedded packages get, the inflate-ratio floor), documented only in prose. Limits::derived() returns them as DerivedLimits, computed by the same functions the reading and writing code calls, so it cannot drift from what is enforced. Limits::get(name) reads a limit by name, and Limits::in_force() returns the limits a call on this thread runs under. office-oxide limits [--strict] [--limit NAME=VALUE] prints both. The C interface has office_options_limits_json, used by ParseOptions.limits() / .derived() (Python), ParseOptions.limits() (Node, exact for values past 2^53, Infinity for no bound), ParseOptions.Limits() (Go), ParseOptions.Limits() (C#) and WasmParseOptions.limits() (WASM, with the Node shapes). ParseOptions::names() lists every option by name, the one list every binding hands out. The README has a limits table (checked against the defaults by a test) and states where options apply: to the document opened with them, while the direct format readers and work on your own threads read the process-wide settings, and ParseOptions::default() copies those when it is made. It also states plainly that there is no time limit, so untrusted input should be read under a process timeout.

Deprecated

  • docx::DocxDocument::images, pptx::PictureShape::data and xlsx::WorksheetPicture::data, which are empty when pictures are read lazily: use DocxDocument::picture / pictures, PptxDocument::slide_picture and XlsxDocument::picture, which work either way. They are removed in 0.2.0, when lazy pictures become the default.

Fixed

  • WASM: WasmParseOptions.names() lists lazy_pictures. The WASM binding built its own list of option names and left out lazy_pictures, which set accepts; every binding now hands out the core's ParseOptions::names(), and each binding's tests check all twelve names.

  • The limits documentation says what text_chars bounds. The README called it the characters of text per document; it bounds the text a document can multiply, a workbook's cell text and the text copied out of shared storage (hyperlink targets, shared headers, notes). A document's own paragraphs grow with the file and are bounded by the size limits.

  • Python: the binding releases the GIL while it parses, converts and writes. Every call into the library kept the GIL for its whole length, waiting for the parse thread while holding it, so no other Python thread ran meanwhile, and a host whose logging handler needed the GIL could block against that thread. Opening, every conversion (plain_text, to_markdown, to_markdown_with_images, to_html, to_ir, to_ir_json, damage), save_as, editing and saving an EditableDocument, the XlsxWriter/PptxWriter save and to_bytes methods, and the module-level functions now run without it, so several Python threads read documents in parallel. Python objects are built after the call returns. A Document, EditableDocument, XlsxWriter or PptxWriter shared between threads takes their calls in turn, as when the GIL serialised them (closing a document while another thread reads it lets that call finish); behaviour is otherwise unchanged.

  • Legacy .ppt: slide text kept in metroBlob packages is now read. A shape can hold its text in an embedded package (drs/shapexml.xml) with the legacy text atoms left nearly empty, as WPS writes an OOXML-origin deck, and such a deck came out almost empty with Ok. A package is used only for a shape with no legacy text of its own, so PowerPoint files that write both don't repeat their text. Titles in packages render as titles, and cached slide-number and date values are left out. Reading is bounded per file and per slide, and text a limit or a damaged package kept back is reported in warnings() and text_complete(). Reported and first implemented by @​xugangqiang.

  • Document::damage() and strict mode classify each warning by the kind it was raised with, not by its words. A loss was told from a repair by looking for "truncat", "limit" or "cut" in the warning text, so a reworded or new warning was misfiled with no test failing, and a document cut in several ways reported one cut. Every reader and converter now records each warning with its kind where it raises it (the warning text is unchanged); damage() reports every cut as Truncated, every repair as Recovered, and a generic "the text is incomplete" entry only when nothing says why: no cut and no unreadable part. A .xls cut by a limit now names the limit instead of that generic entry. A test fails when a converter writes Metadata::warnings without a kind.

Changed (memory)

  • A part is read into a buffer of its declared size, not up to twice it. The buffer for a zip part starts at no more than 1 MiB (the declared size is the file's to state) and doubled from there, so a 2 MiB picture was read into a 4 MiB buffer. It now grows toward the declared size by no more than it holds: a part of the size it declares ends in a buffer of exactly that size, and one that declares more or less costs no more than doubling did. A lazy open of pictures now peaks at one picture rather than two, an eager one at what it keeps plus one.

  • Peak memory is held by a test. tests/test_peak_memory.rs opens and converts thirteen documents built in code (a sparse sheet, many sheets, a long .docx, picture-heavy .docx/.pptx/.xlsx read eagerly and lazily, dense and sparse .xls, a long .doc, a many-slide .ppt) every way the library converts, and fails when the counted heap peak of the open or of the costliest conversion grows by a third. A table test bounds the size of every reader and IR type allocated once per element (docx runs and paragraphs, pptx shapes, xlsx cells and rows, xls cells, doc paragraphs, ppt text runs, IR elements and cells), so a field added inline fails it.

  • A strict open of a .xls or .ppt file no longer converts the document. OnIncomplete::Error (and ParseOptions::strict()) converted the whole document to the IR at open to learn whether it was complete; for these formats the reader alone decides that, so damage() and the strict check read it from the reader. A strict open of a 100,000-cell synthetic workbook peaked at 80 MB, ten times a plain open; it now costs the same as one. Other formats are still converted once.

  • .xlsx: a very large sheet is parsed as it is inflated, never held whole. A sheet part was inflated into one buffer and parsed from it, so opening a workbook held the sheet's whole XML beside the cells built from it: 94 to 313 MiB of XML in the corpus's largest workbooks. A sheet part of 64 MiB or more is now inflated a megabyte at a time and parsed in pieces of about 4 MiB that each end after a complete row. Each piece is parsed after open tags for the elements still open where the last one ended (found by a scan that tokenizes as the XML reader does, run alongside the parse), so the reader is in the state a whole-part parse would be in; rows without r=, shared formulas and everything after the rows come out the same. A part that cannot be read this way (not UTF-8, a checksum or length the whole-part read tolerates, malformed XML, a split that would fall inside markup) is read whole as before, so salvage, truncation checks and reports are unchanged. The work of streaming is bounded against crafted parts: the search for a row end looks at each chunk once (a 250 MB stretch with no row end took 64 s when it was searched again with every chunk), and a piece ends only where the open elements are the root and sheetData (a </row> nested in a row or in headerFooter does not end one); pending XML past 16 MiB without a row end, more than 4 KiB of open tags to carry, or a document type declaration (which package XML may not use) sends the part to the whole-part read. Peak resident memory of office-oxide text falls from 241 to 153 MiB, 343 to 226 MiB and 739 to 421 MiB on three such workbooks, wall time is unchanged, CPU time rises by up to a sixth (the scan), and the output is byte for byte the same.

  • .xlsx: a workbook's sheets are read and parsed in batches, so their XML is not all held at once. Every sheet's part was read before any was parsed (serially or in parallel), so a workbook of many large sheets held the XML of all of them at its peak. Sheets are now read in batches of at most 32 MiB of XML (and at least one sheet), each batch parsed and its XML dropped before the next is read. Opening six sheets of 16 MiB of XML peaked 100 MB above the input; it now peaks under 64 MB. Results, sheet order and shared-formula resolution are unchanged.

  • Legacy .xls: opening a large sheet no longer holds its cells twice. A sheet's cells were gathered into a list (32 bytes each) and the grid was built while the whole list was still alive, so the open peaked at the list, the grid and the Workbook stream together. The grid now takes the cells off the list as it fills and the list shrinks as it empties. The heap peak of opening falls from 144 to 115 MiB, 90 to 73 MiB and 117 to 82 MiB on three 13-28 MB workbooks, with output unchanged on 1,180 corpus files. A cell written by two records still holds the later one.


Install

Rust   cargo add office_oxide

Python   pip install office-oxide

... (truncated)

Changelog

Sourced from office_oxide's changelog.

[0.1.17] - 2026-11-10

Added

  • Pictures can be held as their stored streams, not inflated at open (lazy_pictures, opt-in). Opening a .docx, .pptx or .xlsx inflated every picture and kept it for the life of the document, although text, Markdown and HTML without images never read them: a 22 MB .docx of 39 page images held 341 MiB after open. With the new ParseOptions::lazy_pictures (or read_options::set_lazy_pictures, --limit lazy_pictures=1, or the option by name in every binding), a deflated picture that inflated to at least twice its stored size is kept as that stream and inflated only when it is used. On that file plain_text() peaks at 84 MiB instead of 404, the IR, to_html() and the IR JSON at about 395 instead of 696, and Markdown with images at 873 instead of 1,175, with every output byte for byte the same. Each picture is still read and checked at open, under the same limits and with the same reports, and only one is inflated at a time. The IR still copies each picture into every image that draws it, in both modes; sharing them, which needs a new field on ir::Image, waits for 0.2.0. New API: core::picture::PictureBytes (a shared, cheaply cloned handle with len, bytes and ptr_eq), Image::bytes(), Image::byte_len(), Image::has_bytes(), DocxDocument::picture / pictures, PptxDocument::slide_picture and XlsxDocument::picture. Off by default; without it nothing changes.

  • Large output without a host-language string, in every binding. A Python str is four bytes per character once one character is outside the Basic Multilingual Plane, a .NET string and most JS strings two, so the 450 MB IR JSON of a picture-heavy .docx cost a Python process 2.8 GiB. New forms hand over UTF-8 bytes or stream to a file: in Python plain_text_bytes(), to_markdown_bytes(), to_markdown_with_images_bytes(), to_html_bytes(), to_ir_json_bytes() and write_text / write_markdown(images=...) / write_html / write_ir_json, each to a path or a binary file object (1.5 GiB and 0.7 GiB on that file; what a file object's write raises is raised as itself, and a short write is completed); in Node doc.toBuffer(kind) and doc.writeTo(path, kind); in Go doc.Bytes(kind) and doc.WriteFile(path, kind); in C# doc.ToUtf8Bytes(kind), doc.WriteTo(Stream, kind) and doc.WriteTo(path, kind); in WASM doc.toBuffer(kind) (a Uint8Array); in C office_document_render_bytes and office_document_write_to_path. The output is byte-identical to the string forms encoded as UTF-8; the string methods are unchanged.

  • Reading time is checked against input size, in the tests and in the fuzzer. The limits bound sizes and counts, not time, and nothing caught the next algorithm that grows faster than its input. A new test suite (tests/test_time_scaling.rs) reads each shape that was once quadratic (a wide docx table, unclosed Markdown openers, a replacement across many runs, many clashing sheet names, many sheets, a large .xls, a long .doc) at two sizes through the public API and fails when four times the input costs ten times the time (a quadratic path costs sixteen). The fuzz jobs now fail an input that takes more than 10 s (-timeout=10; libFuzzer's default was 20 minutes). A cooperative step budget was prototyped in the worksheet event loop: its cost was within the noise of the timing, but a counter on parse events bounds only work linear in the input, which the size limits already bound, so it was not added as a limit. The README states that there is no time limit and that untrusted input should be read under a process timeout.

  • What a limit does, before a document is read: Limits::derived(), office-oxide limits and the options objects. The library derives eleven bounds from three of the eight limits by fixed fractions (picture and padding budgets, the write estimate, the share of package_bytes embedded packages get, the inflate-ratio floor), documented only in prose. Limits::derived() returns them as DerivedLimits, computed by the same functions the reading and writing code calls, so it cannot drift from what is enforced. Limits::get(name) reads a limit by name, and Limits::in_force() returns the limits a call on this thread runs under. office-oxide limits [--strict] [--limit NAME=VALUE] prints both. The C interface has office_options_limits_json, used by ParseOptions.limits() / .derived() (Python), ParseOptions.limits() (Node, exact for values past 2^53, Infinity for no bound), ParseOptions.Limits() (Go), ParseOptions.Limits() (C#) and WasmParseOptions.limits() (WASM, with the Node shapes). ParseOptions::names() lists every option by name, the one list every binding hands out. The README has a limits table (checked against the defaults by a test) and states where options apply: to the document opened with them, while the direct format readers and work on your own threads read the process-wide settings, and ParseOptions::default() copies those when it is made. It also states plainly that there is no time limit, so untrusted input should be read under a process timeout.

Deprecated

  • docx::DocxDocument::images, pptx::PictureShape::data and xlsx::WorksheetPicture::data, which are empty when pictures are read lazily: use DocxDocument::picture / pictures, PptxDocument::slide_picture and XlsxDocument::picture, which work either way. They are removed in 0.2.0, when lazy pictures become the default.

Fixed

  • WASM: WasmParseOptions.names() lists lazy_pictures. The WASM binding built its own list of option names and left out lazy_pictures, which set accepts; every binding now hands out the core's ParseOptions::names(), and each binding's tests check all twelve names.

  • The limits documentation says what text_chars bounds. The README called it the characters of text per document; it bounds the text a document can multiply, a workbook's cell text and the text copied out of shared storage (hyperlink targets, shared headers, notes). A document's own paragraphs grow with the file and are bounded by the size limits.

  • Python: the binding releases the GIL while it parses, converts and writes. Every call into the library kept the GIL for its whole length, waiting for the parse thread while holding it, so no other Python thread ran meanwhile, and a host whose logging handler needed the GIL could block against that thread. Opening, every conversion (plain_text, to_markdown, to_markdown_with_images, to_html, to_ir, to_ir_json, damage), save_as, editing and saving an EditableDocument, the XlsxWriter/PptxWriter save and to_bytes methods, and the module-level functions now run without it, so several Python threads read documents in parallel. Python objects are built after the call returns. A Document, EditableDocument, XlsxWriter or PptxWriter shared between threads takes their calls in turn, as when the GIL serialised them (closing a document while another thread reads it lets that call finish); behaviour is otherwise unchanged.

  • Legacy .ppt: slide text kept in metroBlob packages is now read. A shape can hold its text in an embedded package (drs/shapexml.xml) with the legacy text atoms left nearly empty, as WPS writes an OOXML-origin deck, and such a deck came out almost empty with Ok. A package is used only for a shape with no legacy text of its own, so PowerPoint files that write both don't repeat their text. Titles in packages render as titles, and cached slide-number and date values are left out. Reading is bounded per file and per slide, and text a limit or a damaged package kept back is reported in warnings() and text_complete(). Reported and first implemented by @​xugangqiang.

  • Document::damage() and strict mode classify each warning by the kind it was raised with, not by its words. A loss was told from a repair by looking for "truncat", "limit" or "cut" in the warning text, so a reworded or new warning was misfiled with no test failing, and a document cut in several ways reported one cut. Every reader and converter now records each warning with its kind where it raises it (the warning text is unchanged); damage() reports every cut as Truncated, every repair as Recovered, and a generic "the text is incomplete" entry only when nothing says why: no cut and no unreadable part. A .xls cut by a limit now names the limit instead of that generic entry. A test fails when a converter writes Metadata::warnings without a kind.

Changed (memory)

  • A part is read into a buffer of its declared size, not up to twice it. The buffer for a zip part starts at no more than 1 MiB (the declared size is the file's to state) and doubled from there, so a 2 MiB picture was read into a 4 MiB buffer. It now grows toward the declared size by no more than it holds: a part of the size it declares ends in a buffer of exactly that size, and one that declares more or less costs no more than doubling did. A lazy open of pictures now peaks at one picture rather than two, an eager one at what it keeps plus one.

  • Peak memory is held by a test. tests/test_peak_memory.rs opens and converts thirteen documents built in code (a sparse sheet, many sheets, a long .docx, picture-heavy .docx/.pptx/.xlsx read eagerly and lazily, dense and sparse .xls, a long .doc, a many-slide .ppt) every way the library converts, and fails when the counted heap peak of the open or of the costliest conversion grows by a third. A table test bounds the size of every reader and IR type allocated once per element (docx runs and paragraphs, pptx shapes, xlsx cells and rows, xls cells, doc paragraphs, ppt text runs, IR elements and cells), so a field added inline fails it.

  • A strict open of a .xls or .ppt file no longer converts the document. OnIncomplete::Error (and ParseOptions::strict()) converted the whole document to the IR at open to learn whether it was complete; for these formats the reader alone decides that, so damage() and the strict check read it from the reader. A strict open of a 100,000-cell synthetic workbook peaked at 80 MB, ten times a plain open; it now costs the same as one. Other formats are still converted once.

  • .xlsx: a very large sheet is parsed as it is inflated, never held whole. A sheet part was inflated into one buffer and parsed from it, so opening a workbook held the sheet's whole XML beside the cells built from it: 94 to 313 MiB of XML in the corpus's largest workbooks. A sheet part of 64 MiB or more is now inflated a megabyte at a time and parsed in pieces of about 4 MiB that each end after a complete row. Each piece is parsed after open tags for the elements still open where the last one ended (found by a scan that tokenizes as the XML reader does, run alongside the parse), so the reader is in the state a whole-part parse would be in; rows without r=, shared formulas and everything after the rows come out the same. A part that cannot be read this way (not UTF-8, a checksum or length the whole-part read tolerates, malformed XML, a split that would fall inside markup) is read whole as before, so salvage, truncation checks and reports are unchanged. The work of streaming is bounded against crafted parts: the search for a row end looks at each chunk once (a 250 MB stretch with no row end took 64 s when it was searched again with every chunk), and a piece ends only where the open elements are the root and sheetData (a </row> nested in a row or in headerFooter does not end one); pending XML past 16 MiB without a row end, more than 4 KiB of open tags to carry, or a document type declaration (which package XML may not use) sends the part to the whole-part read. Peak resident memory of office-oxide text falls from 241 to 153 MiB, 343 to 226 MiB and 739 to 421 MiB on three such workbooks, wall time is unchanged, CPU time rises by up to a sixth (the scan), and the output is byte for byte the same.

  • .xlsx: a workbook's sheets are read and parsed in batches, so their XML is not all held at once. Every sheet's part was read before any was parsed (serially or in parallel), so a workbook of many large sheets held the XML of all of them at its peak. Sheets are now read in batches of at most 32 MiB of XML (and at least one sheet), each batch parsed and its XML dropped before the next is read. Opening six sheets of 16 MiB of XML peaked 100 MB above the input; it now peaks under 64 MB. Results, sheet order and shared-formula resolution are unchanged.

  • Legacy .xls: opening a large sheet no longer holds its cells twice. A sheet's cells were gathered into a list (32 bytes each) and the grid was built while the whole list was still alive, so the open peaked at the list, the grid and the Workbook stream together. The grid now takes the cells off the list as it fills and the list shrinks as it empties. The heap peak of opening falls from 144 to 115 MiB, 90 to 73 MiB and 117 to 82 MiB on three 13-28 MB workbooks, with output unchanged on 1,180 corpus files. A cell written by two records still holds the later one.

Commits
  • 5f4ce46 Release v0.1.17: lazy and shared pictures, streamed xlsx sheets, typed damage...
  • 7cbd339 docs(examples): one ten-scenario catalogue in every binding, each run under a...
  • 3b0162d fix(ppt): read shape text from metroBlob packages in legacy .ppt files (#529)
  • See full diff in compare view

Updates syn from 3.0.6 to 3.0.10

Release notes

Sourced from syn's releases.

3.0.10

3.0.9

  • Add Punctuated::retain (#2088)

3.0.8

  • Fix parsing of <- and -> tokens in expressions (#2086, #2087)

3.0.7

Commits
  • 97ec37d Release 3.0.10
  • d12a53f Merge pull request #2090 from LucaCappelletti94/upstream/fixup-linear-scan
  • 75cc9dc Update test suite to nightly-2026-11-08
  • cbc91c4 Update test suite to nightly-2026-11-03
  • 38318bf Update test suite to nightly-2026-10-30
  • 4afbff6 Skip rescanning operands whose scan already failed
  • f81c060 Update test suite to nightly-2026-10-27
  • acfda91 Update test suite to nightly-2026-10-26
  • 9ff9287 Update test suite to nightly-2026-10-15
  • 7c44a55 Release 3.0.9
  • Additional commits viewable in compare view

Updates thiserror from 2.0.20 to 2.0.21

Release notes

Sourced from thiserror's releases.

2.0.21

  • Fix parsing of generic unit variants in display expressions (#459)
Commits
  • b1827ee Release 2.0.21
  • 58037b5 Merge pull request #459 from dtolnay/turbofish
  • f82a0cf Keep track of nested turbofish depth
  • 72ea492 Raise required compiler to Rust 1.77
  • 72eea0d Resolve io_other_error clippy lint in tests
  • 07f09a2 Raise required compiler to Rust 1.74
  • 2715388 Update ui test suite to nightly-2026-09-22
  • 5a306c7 Update ui test suite to nightly-2026-09-05
  • ef9383b Update ui test suite to nightly-2026-08-22
  • 8336b84 Update ui tests for version 2.0.20
  • See full diff in compare view

Updates zip from 9.1.3 to 9.5.1

Release notes

Sourced from zip's releases.

v9.5.1

🐛 Bug Fixes

  • (write) raw_copy_file corrupted ZipCrypto metadata (#1288)

v9.5.0

🚀 Features

  • ExtractOptions root_dir filter (#1279)

v9.4.2

🚜 Refactor

  • Replace std::num and std::cmp with core::num and core::cmp (#1276)

⚙️ Miscellaneous Tasks

  • apply cargo fmt to security patches (#1275)

v9.4.1

🐛 Bug Fixes

🚜 Refactor

  • Use boxed slices instead of Vec when growability isn't needed (#1271)

v9.4.0

🚀 Features

  • Implement ZipIterable::files_with_reader (#1267)

v9.3.0

🚀 Features

  • (write) implement TryFrom for PreparedZipFile (#1258)

v9.2.3

🐛 Bug Fixes

  • (ci) resolve typos false-positive, fix upath smuggling test, and add Turkish tests (#1256)

🚜 Refactor

  • Replace std:: with core:: where possible (#1253)

v9.2.2

🚜 Refactor

... (truncated)

Changelog

Sourced from zip's changelog.

9.5.1 - 2026-11-10

🐛 Bug Fixes

  • (write) raw_copy_file corrupted ZipCrypto metadata (#1288)

9.5.0 - 2026-11-08

🚀 Features

  • ExtractOptions root_dir filter (#1279)

9.4.2 - 2026-11-07

🚜 Refactor

  • Replace std::num and std::cmp with core::num and core::cmp (#1276)

9.4.1 - 2026-11-06

🐛 Bug Fixes

🚜 Refactor

  • Use boxed slices instead of Vec when growability isn't needed (#1271)

9.4.0 - 2026-11-05

🚀 Features

  • Implement ZipIterable::files_with_reader (#1267)

9.3.0 - 2026-11-03

🚀 Features

  • (write) implement TryFrom<ZipFileBuilder> for PreparedZipFile (#1258)

9.2.3 - 2026-11-03

🚜 Refactor

  • Replace most equivalent uses of std:: with core:: (#1253)

9.2.2 - 2026-11-01

🚜 Refactor

... (truncated)

Commits
  • ee61677 chore: release v9.5.1 (#1290)
  • d4976d1 fix(write): raw_copy_file keeps an encrypted entry decryptable (#1288)
  • 43297e2 ci(deps): bump the codeql group with 2 updates (#1286)
  • 14ea209 ci(deps): bump github/codeql-action/upload-sarif from 4.39.1 to 4.39.2 (#1287)
  • b03a868 ci(deps): bump rust from 09027f8 to 8a4e90c in /.devcontainer (#1284)
  • 5a8b69a ci(deps): bump release-plz/action from 0.5.147 to 0.5.148 (#1283)
  • 6ec8a2a ci(deps): bump crate-ci/typos from 1.51.1 to 1.52.0 (#1282)
  • ddf3ec3 chore: release v9.5.0 (#1281)
  • 42ad84b feat: ExtractOptions root_dir filter (#1279)
  • ce7aad0 ci: Use an orchestrator pattern to trigger size labels on forked PRs without ...
  • Additional commits viewable in compare view

Updates proptest from 1.11.0 to 1.12.0

Commits
  • 70cf37d Merge pull request #675 from proptest-rs/prep-1.12
  • 729e0ff prepare next releases and fix panic handling
  • 28c1fad Merge pull request #664 from Expyron/main
  • 5ddd9a4 Merge pull request #668 from toddATavail/fix-fork-persisted-failure
  • a6a9ac7 Merge pull request #673 from glaziermag/fix-book-deploy
  • d57d88b Replace rand_chacha with chacha20
  • 9411213 Deploy the book on pushes to main
  • 1719745 fix: terminate the fork replay when a persisted failure fails
  • 0f55b7a test: a persisted failure in a forked test must be reported
  • 4b0594e Merge pull request #669 from proptest-rs/no-fail-fast
  • Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.


Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

  • @dependabot rebase will rebase this PR
  • @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
  • @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
  • @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
  • @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
  • @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
  • @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
  • @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

Bumps the rust-deps group with 6 updates in the / directory:

| Package | From | To |
| --- | --- | --- |
| [serde_json](https://github.com/serde-rs/json) | `1.0.151` | `1.0.154` |
| [office_oxide](https://github.com/yfedoseev/office_oxide) | `0.1.16` | `0.1.17` |
| [syn](https://github.com/dtolnay/syn) | `3.0.6` | `3.0.10` |
| [thiserror](https://github.com/dtolnay/thiserror) | `2.0.20` | `2.0.21` |
| [zip](https://github.com/zip-rs/zip2) | `9.1.3` | `9.5.1` |
| [proptest](https://github.com/proptest-rs/proptest) | `1.11.0` | `1.12.0` |



Updates `serde_json` from 1.0.151 to 1.0.154
- [Release notes](https://github.com/serde-rs/json/releases)
- [Commits](serde-rs/json@v1.0.151...v1.0.154)

Updates `office_oxide` from 0.1.16 to 0.1.17
- [Release notes](https://github.com/yfedoseev/office_oxide/releases)
- [Changelog](https://github.com/yfedoseev/office_oxide/blob/main/CHANGELOG.md)
- [Commits](yfedoseev/office_oxide@v0.1.16...v0.1.17)

Updates `syn` from 3.0.6 to 3.0.10
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](dtolnay/syn@3.0.6...3.0.10)

Updates `thiserror` from 2.0.20 to 2.0.21
- [Release notes](https://github.com/dtolnay/thiserror/releases)
- [Commits](dtolnay/thiserror@2.0.20...2.0.21)

Updates `zip` from 9.1.3 to 9.5.1
- [Release notes](https://github.com/zip-rs/zip2/releases)
- [Changelog](https://github.com/zip-rs/zip2/blob/master/CHANGELOG.md)
- [Commits](zip-rs/zip2@v9.1.3...v9.5.1)

Updates `proptest` from 1.11.0 to 1.12.0
- [Release notes](https://github.com/proptest-rs/proptest/releases)
- [Changelog](https://github.com/proptest-rs/proptest/blob/main/CHANGELOG.md)
- [Commits](proptest-rs/proptest@v1.11.0...v1.12.0)

---
updated-dependencies:
- dependency-name: serde_json
  dependency-version: 1.0.154
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: rust-deps
- dependency-name: office_oxide
  dependency-version: 0.1.17
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: rust-deps
- dependency-name: syn
  dependency-version: 3.0.10
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: rust-deps
- dependency-name: thiserror
  dependency-version: 2.0.21
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: rust-deps
- dependency-name: zip
  dependency-version: 9.5.1
  dependency-type: direct:production
  update-type: version-update:semver-minor
  dependency-group: rust-deps
- dependency-name: proptest
  dependency-version: 1.12.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
  dependency-group: rust-deps
...

Signed-off-by: dependabot[bot] <support@github.com>
@dependabot @github

dependabot Bot commented on behalf of github Nov 16, 2026

Copy link
Copy Markdown
Contributor Author

Labels

The following labels could not be found: dependencies, rust. Please create them before Dependabot can add them to a pull request.

Please fix the above issues or remove invalid values from dependabot.yml.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants