Local guide to Gitignore, SemVer, source maps and WASM imports
Inspect supplied path rules, version constraints, original stack locations and static WASM descriptors. Keep the scope of each report explicit.
Check Gitignore decisions against the intended root
Paste one root .gitignore and case-sensitive relative paths. Mark directories with a trailing slash. The report gives the final matching rule line and an excluded parent when that parent blocks a later negation. For *.log followed by !keep.log, a.log is ignored and keep.log is explicitly included.
A build/ rule followed only by !build/keep.txt still excludes build/keep.txt: the parent directory remains excluded. Supply a rule that also permits traversal when that is the intended policy. This check does not know tracked files, nested .gitignore files, global excludes or whether supplied paths actually exist. Combined rules and paths are limited to 1 MiB, 5,000 rules and 10,000 paths.
| Check | Keep with the report |
|---|---|
| Gitignore | Root scope, path type and blocking parent |
| SemVer | Exact range, prerelease option and original order |
| Source map | Generated filename and input column basis |
| WASM | Browser engine and static descriptor scope |
Use the same SemVer range and prerelease policy
Supply complete strict versions such as 1.2.0+build.2 and an npm range such as ^1.2.0. Blank range means *. The report preserves original strings and source lines, sorts by precedence and shows a null maximum when nothing matches. A list and range have a combined 1 MiB budget and at most 10,000 version lines.
Build metadata does not change precedence. Equal versions retain input order, so 1.2.0+build.2 and 1.2.0+build.1 remain tied; the first matching input wins maximum selection. The default npm policy excludes many prereleases unless the range admits their tuple. Turn on the visible prerelease option only when that wider policy is intended. Registry dist-tags and actual package availability are outside this supplied-list check.
Make generated files and column numbering explicit
Supply an ordinary v3 source map plus the generated filename and a stack or coordinate list. Stack columns are often displayed from one, while source-map original columns are zero-based. Choose the input column basis explicitly and keep it beside the mapped text. Wrong filenames, unmapped positions and unparsed lines remain visible in the complete report.
The combined map and frame budget is 5 MiB with at most 10,000 frames. Indexed section maps and malformed mappings are rejected. The mapper resolves recorded source paths and optional names locally, but does not fetch source URLs. Mapping a location is different from recovering a source file: absent sourcesContent cannot be reconstructed from coordinates.
Inspect WASM host names without running the module
Select a core .wasm binary up to 1 MiB. The current browser validates and compiles a Module, then lists import module names, import/export names and descriptor kinds. It never creates an Instance or calls an import, start function or exported function. The sample module provides a small repeatable import/export example.
A descriptor kind does not reveal a full function signature or prove that a host implementation is compatible. Empty names, Unicode names and zero-length descriptor lists are valid cases. Compilation rejection can mean damaged bytes or a feature unsupported by the current engine; use a compatible inspection environment before concluding that the binary is universally invalid.
Retain the complete output and recover deliberately
The visible tables preview at most 200 records. Download complete JSON or CSV for larger sets; source-map mapping also provides copyable mapped text. CSV spreadsheet protection may prefix dangerous cell values, while JSON retains original strings.
On an error, correct the reported path, version, mapping or binary rather than treating a partial table as a successful result. Keep the original file and input settings; a different root, range, filename or column basis can produce a different valid answer.
Before handing over the report
- Retain original inputs and the explicit settings.
- Check unmatched, unmapped and unavailable records.
- Download complete records when the table is truncated.
- Verify the receiving workflow uses the same root, range, coordinates or host contract.
References
- .gitignore 规则测试
Specific first-person task; the current tool states its supported scope.
- SemVer 范围测试
Specific first-person task; the current tool states its supported scope.
- Source map 堆栈还原
Specific first-person task; the current tool states its supported scope.
- WebAssembly 导入导出检查
Specific first-person task; the current tool states its supported scope.
Tools in this category
Expand a tool to see its steps, options and supported formats, then open its workspace.
Gitignore path inspectorTest supplied paths under one root Gitignore file and trace the final matching rule and excluded parent directory.
Explain an unexpected ignore decision before changing your rules. Supply the exact path spelling, mark directories with a trailing slash, and inspect which rule wins.
Steps
- Paste or open one root .gitignore and enter one path per line. Keep path case and meaningful spaces.
- Use a trailing slash for directories, then inspect status, rule line and any blocking parent.
- Download the full report. Check tracked/global/nested rules separately in the actual repository.
Available options
- Protect CSV cells in spreadsheets
- On by default
Capabilities and limits
- One UTF-8 rules file or pasted rules and a path list, with a combined 1 MiB budget. Up to 5,000 physical rule lines and 10,000 paths, each at most 1,024 UTF-16 units and 32 components. Matching stops at 2,000,000 rule tests. Selected file takes precedence.
- Paths are case-sensitive, relative POSIX names under the supplied root rules. A trailing slash declares a directory; an unmarked final component declares a file. No absolute paths, backslashes, empty components, . or .. components, or control characters. No repository, tracked-file, global-ignore or nested-ignore lookup.
- Uses the patterns compiled by locked ignore 7.0.5, including escaped #/!, spaces, character classes, ** and negation. The last directly matching rule wins. An excluded parent blocks descendant re-inclusion; its winning rule and path are reported. Empty/comment-only rules are allowed. LF and CRLF are supported; CR-only is rejected.
- The table shows the first 200 supplied paths; JSON and CSV include all paths. Formula-like CSV cells receive an apostrophe by default. JSON preserves exact pattern/path strings.
SemVer range version selectorCheck strict supplied versions against an npm SemVer range, sort by precedence and find the highest matching version.
Review a dependency constraint against a known version list. See which releases match, how prerelease inclusion changes the result, and which original version string is selected.
Steps
- Enter one strict version per line and the exact npm range.
- Choose whether to include prereleases, then inspect matching rows and the selected maximum.
- Download the report and confirm the receiving package manager uses the same range/prerelease policy.
Available options
- Include prereleases
- Off by default
- Protect CSV cells in spreadsheets
- On by default
Capabilities and limits
- One UTF-8 version-list file or pasted list plus a range, with a combined 1 MiB budget. At most 10,000 nonempty version lines, 256 characters per version and 4,096 per range. Blank lines are skipped without changing reported source line numbers. Selected file takes precedence.
- Versions must use strict major.minor.patch syntax with optional prerelease/build metadata; v/= prefixes, leading zeros and surrounding spaces are rejected. Major/minor/patch and numeric prerelease identifiers must fit JavaScript safe integers. Larger numeric prerelease identifiers are explicitly unsupported.
- Range syntax follows locked node-semver 7.7.3, including comparator sets, ||, caret, tilde, hyphen and X ranges. An empty range means *. Prereleases use its default exclusion unless the visible include-prerelease option is enabled or the range specifically admits that prerelease tuple. No registry, dist-tag or release-status lookup.
- Build metadata stays in the output but does not affect precedence. Equal-precedence versions keep input order; maximum selection follows maxSatisfying and returns the first equal-precedence input. Preview shows 200 rows; JSON/CSV contain all rows. Spreadsheet protection affects CSV only.
Source map stack mapperMap explicit generated stack frames or coordinates through an ordinary v3 source map, preserving unmapped frames and column bases.
Recover original source positions from a local generated stack. Supply its matching map and exact generated filename, then inspect each mapped, unmapped or other-file frame.
Steps
- Open a local ordinary .map file, or paste its JSON, and enter the exact generated file/URL that appears in the stack.
- Paste a stack or choose coordinate JSON. Set its generated-column basis before mapping.
- Inspect mapped status and original zero-based columns. Download the complete report and retained unmapped text.
Available options
- Exact generated filename / URL
- bundle.js
- Frame input
- Stack text · Coordinate JSON
- Generated column numbering
- One-based · Zero-based
- Protect CSV cells in spreadsheets
- On by default
Capabilities and limits
- One UTF-8 source-map file or pasted map plus stack/coordinate input, combined up to 5 MiB. At most 10,000 frames, 10,000 source/name entries, 100,000 generated mapping lines and 100,000 segments. Selected file takes precedence.
- Accepts ordinary version-3 maps only; indexed sections are rejected. Strictly checks VLQ segment shapes, source/name indices, ordered unique generated columns and nonnegative original positions. VLQ values and positions are bounded to the compatible signed 32-bit decoder range. Duplicate/normalized-colliding source names are rejected.
- Generated lines are one-based. Choose zero-based or one-based generated columns visibly; original columns are always zero-based. Matching uses the explicit generated filename exactly and the greatest-lower-bound segment on the same generated line. No filename guessing, network fetching or source execution. sourceRoot and source names remain visible.
- Supports finite Chrome-style at file:line:column / at name (file:line:column), Firefox name@file:line:column and bare coordinates, or JSON arrays of {file?,line,column}. Unparsed lines, unmapped positions and frames from other files remain in the report. Preview shows 200 rows; complete JSON/CSV and mapped text are downloadable.
WASM import and export inspectorList a local core WebAssembly module’s import/export descriptors without instantiating it or running its start function.
Inspect which imports a WASM module declares and which names it exports before wiring it into an application. Module bytes stay in your browser and its code is not executed.
Steps
- Choose a local binary WASM module, or load the explicitly marked demonstration.
- Inspect import namespaces, exact names and descriptor kinds; check the visible no-instance result.
- Download the report and verify full runtime compatibility separately in your target environment.
Available options
- Use built-in example when no file is selected
- Off by default
- Protect CSV cells in spreadsheets
- On by default
Capabilities and limits
- One binary .wasm file up to 1 MiB. At most 10,000 combined import/export descriptors. The selected file takes precedence over the optional built-in example. Enable the example explicitly to inspect and download its real module bytes.
- Checks the WebAssembly magic header and core version 1, then uses the current browser engine’s validate/Module APIs. A compile rejection can mean corrupted bytes or unsupported engine features; it does not prove invalidity on every engine.
- Reads descriptor direction, import module/name, export name and kind, including empty/Unicode names. An empty valid module produces zero descriptors. Does not instantiate, call imports, execute a start function, reconstruct signatures or inspect code/data/custom sections.
- Module compilation runs inside the existing tool worker with its watchdog. Preview shows 200 descriptors; JSON/CSV contain all. Spreadsheet protection applies only to CSV. A module may be valid yet lack the imports/exports you expected; no runtime behavior is inferred.