Вирт - русифицированный паскаль. A modern, self-hosting Object Pascal compiler built for the 2020s. Zero legacy, full ARC, and unified UTF-8.
Find a file
Graeme Geldenhuys 8122b0a299 Wire semantic analysis pass into the compiler pipeline
The semantic analyser now runs between parsing and code generation.
Type mismatches, undeclared identifiers, and duplicate declarations are
reported to stderr with source position and exit code 1 before any IR
is emitted.
2026-04-20 18:01:03 +01:00
compiler Wire semantic analysis pass into the compiler pipeline 2026-04-20 18:01:03 +01:00
docs Phase 1 bootstrap: Blaise compiler skeleton 2026-04-20 15:35:50 +01:00
rtl Phase 1 bootstrap: Blaise compiler skeleton 2026-04-20 15:35:50 +01:00
tools Phase 1 bootstrap: Blaise compiler skeleton 2026-04-20 15:35:50 +01:00
vendor/qbe Fix compilation errors; vendor QBE 1.2; end-to-end Hello World works 2026-04-20 16:38:57 +01:00
.gitignore Fix compilation errors; vendor QBE 1.2; end-to-end Hello World works 2026-04-20 16:38:57 +01:00
LICENCE Initial project scaffold 2026-04-20 14:22:10 +01:00
project.xml Phase 1 bootstrap: Blaise compiler skeleton 2026-04-20 15:35:50 +01:00
README.adoc Phase 1 bootstrap: Blaise compiler skeleton 2026-04-20 15:35:50 +01:00

= Blaise
:toc: left
:toclevels: 2
:icons: font
:source-highlighter: rouge

A modern, cross-platform Object Pascal compiler targeting native code via
https://c9x.me/compile/[QBE] (and eventually LLVM). Single language mode,
single string type, zero-GUID interfaces, reified generics, and first-class
https://github.com/graemeg/opdebugger[OPDF] debug format support.

Bootstrapped with https://www.freepascal.org/[Free Pascal Compiler].
Built with https://github.com/graemeg/pasbuild[PasBuild].
Licensed under the BSD 3-Clause licence.

== Design Philosophy

The Object Pascal ecosystem has two options: Embarcadero Delphi (proprietary,
Windows-first) and Free Pascal (open source but carrying 30 years of accumulated
complexity — five language modes, five string types, and thousands of include files).

This compiler takes a different approach:

* *One language mode.* No `{$mode}` switches; no legacy dialect support.
* *One string type.* UTF-8 reference-counted string. `RawBytes` for binary data.
* *Clean interfaces.* No COM GUIDs; interface dispatch via compile-time vtable mapping.
* *Reified generics.* Monomorphization at compile time — no type erasure.
* *Modern build system.* PasBuild with `project.xml`; no makefiles.
* *First-class debugger.* OPDF is the default debug format; DWARF is not required.

See link:docs/design.adoc[docs/design.adoc] for the full architecture and
implementation plan.

== Repository Layout

This project uses PasBuild's multi-module layout. Each subdirectory with a
`project.xml` is an independent module; the root `project.xml` is the aggregator.

....
project.xml                       Root aggregator (packaging=pom)
│
├── compiler/                     The compiler binary (packaging=application)
│   ├── project.xml
│   └── src/
│       ├── main/pascal/          uLexer, uParser, uAST, uCodeGenQBE, ...
│       └── test/pascal/          FPTest test suite for compiler units
│
├── rtl/                          Runtime library (packaging=library)
│   ├── project.xml
│   └── src/
│       ├── main/pascal/          System.pas, SysUtils.pas, Classes.pas, ...
│       └── test/pascal/          FPTest test suite for RTL units
│
├── tools/
│   └── migration-analyser/       FPC/Delphi migration report tool (packaging=application)
│       ├── project.xml           depends on compiler module
│       └── src/
│           ├── main/pascal/
│           └── test/pascal/
│
├── vendor/qbe/                   Vendored QBE backend source (pinned, built from source)
└── docs/                         Design documents and specifications
....

PasBuild compiles each module to its own `target/` subdirectory. Build output is
never committed to the repository.

== Status

[cols="1,3,1", options="header"]
|===
| Phase | Goal | Status

| 1
| Bootstrap pipeline — Hello World on Linux x86_64 via PasBuild
| In progress

| 2
| Type system — classes, records, ARC, exceptions
| Planned

| 3
| Generics + zero-GUID interfaces
| Planned

| 4
| OPDF debug info emission
| Planned

| 5
| Self-hosting + LLVM + Windows
| Planned

| 6
| LSP + VS Code extension
| Planned

| 7
| Migration analyser for FPC/Delphi codebases
| Planned
|===

== Building

=== Prerequisites

* Free Pascal Compiler 3.2.2 or later (stable; 3.3.x development snapshots are not required)
* https://github.com/graemeg/pasbuild[PasBuild]
* A C compiler (`gcc` or `clang`) for building the vendored QBE backend
* GNU `ld` or `lld` (Linux); `ld` (macOS)

=== Build all modules

[source,shell]
----
pasbuild compile
----

PasBuild resolves the module dependency order automatically and compiles
`rtl` → `compiler` → `tools/migration-analyser`.

=== Build with a profile

[source,shell]
----
pasbuild compile -p debug      # includes -g -gl -Criot -gh
pasbuild compile -p release    # includes -O2 -CX -XX -Xs
----

=== Run tests

[source,shell]
----
pasbuild test
----

=== Build a single module

[source,shell]
----
pasbuild compile -m blaise-compiler
----

=== Running the compiler

Once built, the compiler binary is at `compiler/target/blaise`.

[source,shell]
----
# Compile a single file
compiler/target/blaise --source Hello.pas --output Hello

# Compile via project.xml
compiler/target/blaise --project project.xml --config debug --output myapp

# Emit QBE IR (useful for debugging the compiler itself)
compiler/target/blaise --source Hello.pas --emit-ir
----

== What Is Dropped From Classic Pascal

[cols="1,3", options="header"]
|===
| Feature | Reason for removal

| `ShortString`, `AnsiString`, `WideString`, `UnicodeString`
| Replaced by a single UTF-8 reference-counted `string` type

| `with` statement
| Source of hard-to-diagnose symbol resolution bugs; breaks static analysis

| Old-style `object` types
| Use `record` (stack/value) or `class` (heap/reference) instead

| COM-style interface GUIDs
| Interface dispatch via compile-time vtable; GUIDs are unnecessary complexity

| Multiple language modes
| One dialect, maintained well, beats five dialects maintained poorly

| `assign`, `reset`, `rewrite`, `blockread`
| Replaced by a stream-based I/O RTL
|===

== Licence

BSD 3-Clause. See link:LICENCE[LICENCE].