Language support for IFC files (ISO 10303-21 "STEP Part 21" text), backed by a native
language server written in Rust. Fast enough for real models: a 74 MB file indexes in well
under a second and edits are re-indexed incrementally.
Features
Syntax highlighting for IFC/STEP files and EXPRESS schemas.
Go to definition and find references on #123 instance references.
Go to schema: F12 on an entity name, typed value or enumeration value opens its EXPRESS
declaration.
Hover on an entity constructor shows its attributes, including inherited ones, next to the
actual arguments.
Diagnostics for obviously invalid content: unknown or abstract entities, wrong argument
counts, undefined or duplicate ids, bad enumeration values, type mismatches and syntax
errors. Full conformance checking (WHERE rules) is out of scope.
Schemas embedded: IFC2X3, IFC4, IFC4X1, IFC4X2, IFC4X3_ADD1 and IFC4X3_ADD2, chosen from
the file's FILE_SCHEMA header. Extra .exp files can be added through settings.
Large files
VS Code itself limits what extensions can do with large documents:
File size
What works
≤ 20 MB
everything
20–50 MB
everything except syntax colouring, which VS Code disables for large files unless editor.largeFileOptimizations is false
> 50 MB
nothing: VS Code does not share documents above 50 MB with any extension, so the language server never sees them
Settings
Setting
Default
Meaning
ifc.serverPath
""
Path to an ifc-lsp binary; empty uses the one bundled here
ifc.extraSchemaDirectories
[]
Directories with additional .exp schemas
ifc.diagnostics.maxProblems
1000
Cap on reported problems per file
ifc.trace.server
off
LSP message tracing
Licence
Apache-2.0. The embedded EXPRESS schemas are buildingSMART International's and are
redistributed unmodified under their own terms. Source: https://github.com/JR-Morgan/IFC-LSP.
