MSBuild Structured Log Viewer search query syntax ================================================== A query is a list of space-separated terms. All terms must match (AND). Matching is substring and case-insensitive by default. Use double quotes for literal/exact matching: Copying file substring match for the two words 'copying' AND 'file' "Copying file" literal substring match for the exact phrase "Copy" exact (whole-string, case-insensitive) match — quotes around a single word turn off substring search Node-kind filters ----------------- Prefix a `$kind` token to restrict results to a particular node type: $project Project nodes $projectevaluation Project evaluation nodes $target Target nodes $task Task invocations $error Errors $warning Warnings $message Messages $property Property nodes $item Item nodes $additem AddItem nodes $removeitem RemoveItem nodes $metadata Metadata nodes $import Import nodes (resolved imports) $noimport NoImport nodes (failed/skipped imports) $secret Detected secrets (if redaction is enabled) Aliases (expanded internally): $csc → $task csc $rar → $task ResolveAssemblyReference Find a node by its index (TimedNode.Index): $42 the node with Index 42 Find by evaluation id ---------------------------------------------------------- ProjectEvaluation titles end with `id:N`, where N is the evaluation id. Project execution nodes cross-reference their evaluation in a message that contains `id:N`. So you can correlate a Project execution with its ProjectEvaluation (and vice versa) by searching for the literal `id:N`: id:29 every node whose text contains 'id:29' — finds both the ProjectEvaluation with that id AND the Project execution(s) that reference it. Combine with `$projectevaluation` to get just the evaluation node: $projectevaluation id:29 Field filters ------------- For NameValue-style nodes (Property, Metadata, Item): name=Configuration name contains 'Configuration' (substring) name="Configuration" name equals 'Configuration' exactly value=Debug value contains 'Debug' name=Foo value=Bar both filters apply Other field filters: skipped=true only skipped targets skipped=false exclude skipped targets height=0 projects with no project references height=N projects whose deepest reference chain is N height=max tallest project(s); also reports the max value Hierarchical scoping -------------------- under(QUERY) match only nodes whose ancestor chain contains a node matching QUERY notunder(QUERY) exclude nodes whose ancestor chain contains a match for QUERY project(QUERY) like under(), but only matches the NEAREST parent project not(QUERY) exclude nodes that themselves match QUERY QUERY may be any nested query, including more under()/notunder()/project()/not(). Multiple under() / project() / notunder() clauses are unioned. Examples: $csc under($project Core) Csc invocations anywhere under a project named *Core* $error project(MyApp) errors whose nearest parent project matches MyApp $warning notunder($task Csc) warnings not produced by a Csc task $rar under(A) notunder(B) notunder(Evaluation) RAR tasks under A but not under B and not under any Evaluation $metadata name=CopyToOutputDirectory value=PreserveNewest under($item .txt under($additem None)) Time intervals (filter) ----------------------- Filter Project / Target / Task results by start or end timestamp. Timestamp must be quoted; any DateTime.Parse-able format works. start<"2024-01-15 14:30:00" started before that timestamp start>"2024-01-15 14:30:00" started after end<"2024-01-15 14:30:00.500" ended before end>"2024-01-15 14:30:00.500" ended after Time annotations (display + sort) --------------------------------- Append these flags to include extra fields in results and sort accordingly. They are NOT filters — they don't reduce the result set. $time include duration; sort by duration descending (alias: $duration) $start include start time; sort by start time (alias: $starttime) $end include end time; sort by end time (alias: $endtime) Examples: $task $time all tasks, slowest first $target $time under($project MyApp) targets in MyApp, slowest first $project $start projects in start-time order Specialized indexes ------------------- $copy — file-copy map $copy filename all copies whose source or dest contains 'filename' $copy directory\path all copies in or out of that directory $copy C:\full\file\path all copies involving the specific file $nuget — NuGet / project.assets.json $nuget project(MyProject.csproj) list MyProject's NuGet dependencies $nuget project(MyProject.csproj) Pkg.Name search dependencies, resolved packages and files $nuget project(.csproj) 13.0.3 search across all projects (slow); version-aware $nuget project(MyProject.csproj) File.dll find a file from a NuGet package $projectreference — project reference graph $projectreference project(MyProject.csproj) RefProj projects referenced by MyProject.csproj directly or indirectly (and matching RefProj). When the scope narrows to a single project, the projects REFERENCING it are also shown — bidirectional. $projectreference project(App.csproj) transitive closure referenced by App.csproj $projectreference project(App.csproj) Core filter to references whose name contains Core $project height=N projects at a given level in the reference graph Tips ---- * The query parser is whitespace-tolerant: `under (` and `under(` both work. Same for `not (`, `notunder (`, `project (`, `start <`, `name =`, etc. * Substring matches are very broad; use $kind filters to narrow scope first. * Use `under()` / `notunder()` instead of trying to express scope via free text. * Quoted single words ("Csc") give whole-string matches — useful when a short token would otherwise collide with longer strings. * `$NN` (e.g. `$1234`) returns the single node with that TimedNode.Index.