Internals of the Nim Compiler

Source   Edit  

Author:Andreas Rumpf
Version:2.0.2

"Abstraction is layering ignorance on top of reality." -- Richard Gabriel

Directory structure

The Nim project's directory structure is:

PathPurpose
bingenerated binary files
buildgenerated C code for the installation
compilerthe Nim compiler itself; note that this code has been translated from a bootstrapping version written in Pascal, so the code is not a poster child of good Nim code
configconfiguration files for Nim
distadditional packages for the distribution
docthe documentation; it is a bunch of reStructuredText files
libthe Nim library

Bootstrapping the compiler

Note: Add . to your PATH so that koch can be used without the ./.

Compiling the compiler is a simple matter of running:

nim c koch.nim
koch boot -d:release

For a debug version use:

nim c koch.nim
koch boot

And for a debug version compatible with GDB:

nim c koch.nim
koch boot --debuginfo --linedir:on

The koch program is Nim's maintenance script. It is a replacement for make and shell scripting with the advantage that it is much more portable. More information about its options can be found in the koch documentation.

Reproducible builds

Set the compilation timestamp with the SOURCE_DATE_EPOCH environment variable.

export SOURCE_DATE_EPOCH=$(git log -n 1 --format=%at)
koch boot # or `./build_all.sh`

Debugging the compiler

Bisecting for regressions

There are often times when there is a bug that is caused by a regression in the compiler or stdlib. Bisecting the Nim repo commits is a useful tool to identify what commit introduced the regression.

Even if it's not known whether a bug is caused by a regression, bisection can reduce debugging time by ruling it out. If the bug is found to be a regression, then you focus on the changes introduced by that one specific commit.

koch temp returns 125 as the exit code in case the compiler compilation fails. This exit code tells git bisect to skip the current commit:

git bisect start bad-commit good-commit
git bisect run ./koch temp -r c test-source.nim

You can also bisect using custom options to build the compiler, for example if you don't need a debug version of the compiler (which runs slower), you can replace ./koch temp by explicit compilation command, see Bootstrapping the compiler.

Building an instrumented compiler

Considering that a useful method of debugging the compiler is inserting debug logging, or changing code and then observing the outcome of a testcase, it is fastest to build a compiler that is instrumented for debugging from an existing release build. koch temp provides a convenient method of doing just that.

By default, running koch temp will build a lean version of the compiler with -d:debug enabled. The compiler is written to bin/nim_temp by default. A lean version of the compiler lacks JS and documentation generation.

bin/nim_temp can be directly used to run testcases, or used with testament with testament --nim:bin/nim_temp r tests/category/tsometest.

koch temp will build the temporary compiler with the -d:debug enabled. Here are compiler options that are of interest when debugging:

  • -d:debug: enables assert statements and stacktraces and all runtime checks
  • --opt:speed: build with optimizations enabled
  • --debugger:native: enables --debuginfo --lineDir:on for using a native debugger like GDB, LLDB or CDB
  • -d:nimDebug cause calls to quit to raise an assertion exception
  • -d:nimDebugUtils: enables various debugging utilities; see compiler/debugutils
  • -d:stacktraceMsgs -d:nimCompilerStacktraceHints: adds some additional stacktrace hints; see https://github.com/nim-lang/Nim/pull/13351
  • -u:leanCompiler: enable JS and doc generation

Another method to build and run the compiler is directly through koch:

koch temp [options] c test.nim

# (will build with js support)
koch temp [options] js test.nim

# (will build with doc support)
koch temp [options] doc test.nim

Debug logging

"Printf debugging" is still the most appropriate way to debug many problems arising in compiler development. The typical usage of breakpoints to debug the code is often less practical, because almost all code paths in the compiler will be executed hundreds of times before a particular section of the tested program is reached where the newly developed code must be activated.

To work around this problem, you'll typically introduce an if statement in the compiler code detecting more precisely the conditions where the tested feature is being used. One very common way to achieve this is to use the mdbg condition, which will be true only in contexts, processing expressions and statements from the currently compiled main module:

# inside some compiler module
if mdbg:
  debug someAstNode

Using the isCompilerDebug condition along with inserting some statements into the testcase provides more granular logging:

# compilermodule.nim
if isCompilerDebug():
  debug someAstNode

# testcase.nim
proc main =
  {.define(nimCompilerDebug).}
  let a = 2.5 * 3
  {.undef(nimCompilerDebug).}

Logging can also be scoped to a specific filename as well. This will of course match against every module with that name.

if `??`(conf, n.info, "module.nim"):
  debug(n)

The above examples also makes use of the debug proc, which is able to print a human-readable form of an arbitrary AST tree. Other common ways to print information about the internal compiler types include:

# pretty print PNode

# pretty prints the Nim ast
echo renderTree(someNode)

# pretty prints the Nim ast, but annotates symbol IDs
echo renderTree(someNode, {renderIds})

# pretty print ast as JSON
debug(someNode)

# print as YAML
echo treeToYaml(config, someNode)


# pretty print PType

# print type name
echo typeToString(someType)

# pretty print as JSON
debug(someType)

# print as YAML
echo typeToYaml(config, someType)


# pretty print PSym

# print the symbol's name
echo symbol.name.s

# pretty print as JSON
debug(symbol)

# print as YAML
echo symToYaml(config, symbol)


# pretty print TLineInfo
lineInfoToStr(lineInfo)


# print the structure of any type
repr(someVar)

Here are some other helpful utilities:

# how did execution reach this location?
writeStackTrace()

These procs may not already be imported by the module you're editing. You can import them directly for debugging:

from astalgo import debug
from types import typeToString
from renderer import renderTree
from msgs import `??`

Native debugging

Stepping through the compiler with a native debugger is a very powerful tool to both learn and debug it. However, there is still the need to constrain when breakpoints are triggered. The same methods as in Debug logging can be applied here when combined with calls to the debug helpers enteringDebugSection() and exitingDebugSection().

  1. Compile the temp compiler with --debugger:native -d:nimDebugUtils
  2. Set your desired breakpoints or watchpoints.
  3. Configure your debugger:
    • GDB: execute source tools/compiler.gdb at startup
    • LLDB execute command source tools/compiler.lldb at startup
  4. Use one of the scoping helpers like so:

if isCompilerDebug():
  enteringDebugSection()
else:
  exitingDebugSection()

A caveat of this method is that all breakpoints and watchpoints are enabled or disabled. Also, due to a bug, only breakpoints can be constrained for LLDB.

The compiler's architecture

Nim uses the classic compiler architecture: A lexer/scanner feeds tokens to a parser. The parser builds a syntax tree that is used by the code generators. This syntax tree is the interface between the parser and the code generator. It is essential to understand most of the compiler's code.

Semantic analysis is separated from parsing.

Short description of Nim's modules

ModuleDescription
nimmain module: parses the command line and calls main.MainCommand
mainimplements the top-level command dispatching
nimconfimplements the config file reader
syntaxesdispatcher for the different parsers and filters
filter_tmplstandard template filter (#? stdtempl)
lexbasebuffer handling of the lexical analyser
lexerlexical analyser
parserNim's parser
rendererNim code renderer (AST back to its textual form)
optionscontains global and local compiler options
asttype definitions of the abstract syntax tree (AST) and node constructors
astalgoalgorithms for containers of AST nodes; converting the AST to YAML; the symbol table
passesimplement the passes manager for passes over the AST
treessome algorithms for nodes; this module is less important
typesmodule for traversing type graphs; also contain several helpers for dealing with types
sigmatchcontains the matching algorithm that is used for proc calls
semexprscontains the semantic checking phase for expressions
semstmtscontains the semantic checking phase for statements
semtypescontains the semantic checking phase for types
seminstinstantiation of generic procs and types
semfoldcontains code to deal with constant folding
sempass2Second semantic checking pass over the AST
vmcontains an AST interpreter for compile time evaluation
pragmassemantic checking of pragmas
identsimplements a general mapping from identifiers to an internal representation (PIdent) that is used so that a simple id-comparison suffices to establish whether two Nim identifiers are equivalent
transftransformations on the AST that need to be done before code generation
cgenmain file of the C code generator
ccgutilscontains helpers for the C code generator
ccgtypesthe generator for C types
ccgstmtsthe generator for statements
ccgexprsthe generator for expressions
extccompthis module calls the C compiler and linker; interesting if you want to add support for a new C compiler

The syntax tree

The syntax tree consists of nodes which may have an arbitrary number of children. Types and symbols are represented by other nodes, because they may contain cycles. The AST changes its shape after semantic checking. This is needed to make life easier for the code generators. See the "ast" module for the type definitions. The macros module contains many examples how the AST represents each syntactic structure.

Runtimes

Nim has two different runtimes, the "old runtime" and the "new runtime". The old runtime supports the old GCs (markAndSweep, refc, Boehm), the new runtime supports ARC/ORC. The new runtime is active when defined(nimV2).

Coding Guidelines

  • We follow Nim's official style guide, see NEP1.
  • Max line length is 100 characters.
  • Provide spaces around binary operators if that enhances readability.
  • Use a space after a colon, but not before it.
  • (deprecated) Start types with a capital T, unless they are pointers/references which start with P.
  • Prefer import package over from package import symbol.

See also the API naming design document.

Porting to new platforms

Porting Nim to a new architecture is pretty easy, since C is the most portable programming language (within certain limits) and Nim generates C code, porting the code generator is not necessary.

POSIX-compliant systems on conventional hardware are usually pretty easy to port: Add the platform to platform (if it is not already listed there), check that the OS, System modules work and recompile Nim.

The only case where things aren't as easy is when old runtime's garbage collectors need some assembler tweaking to work. The default implementation uses C's setjmp function to store all registers on the hardware stack. It may be necessary that the new platform needs to replace this generic code by some assembler code.

Files that may need changed for your platform include:

  • compiler/platform.nim Add os/cpu properties.
  • lib/system.nim Add os/cpu to the documentation for system.hostOS and system.hostCPU.
  • compiler/options.nim Add special os/cpu property checks in isDefined.
  • compiler/installer.ini Add os/cpu to Project.Platforms field.
  • lib/system/platforms.nim Add os/cpu.
  • std/private/osseps.nim Add os specializations.
  • lib/pure/distros.nim Add os, package handler.
  • tools/niminst/makefile.nimf Add os/cpu compiler/linker flags.
  • tools/niminst/buildsh.nimf Add os/cpu compiler/linker flags.

If the --os or --cpu options aren't passed to the compiler, then Nim will determine the current host os, cpu and endianness from system.cpuEndian, system.hostOS and system.hostCPU. Those values are derived from compiler/platform.nim.

In order for the new platform to be bootstrapped from the csources, it must:

  • have compiler/platform.nim updated
  • have compiler/installer.ini updated
  • have tools/niminst/buildsh.nimf updated
  • have tools/niminst/makefile.nimf updated
  • be backported to the Nim version used by the csources
  • the new csources must be pushed
  • the new csources revision must be updated in config/build_config.txt

Runtime type information

Note: This section describes the "old runtime".

Runtime type information (RTTI) is needed for several aspects of the Nim programming language:

Garbage collection
The old GCs use the RTTI for traversing arbitrary Nim types, but usually only the marker field which contains a proc that does the traversal.
Complex assignments
Sequences and strings are implemented as pointers to resizable buffers, but Nim requires copying for assignments. Apart from RTTI the compiler also generates copy procedures as a specialization.

We already know the type information as a graph in the compiler. Thus, we need to serialize this graph as RTTI for C code generation. Look at the file lib/system/hti.nim for more information.

Magics and compilerProcs

The system module contains the part of the RTL which needs support by compiler magic. The C code generator generates the C code for it, just like any other module. However, calls to some procedures like addInt are inserted by the generator. Therefore, there is a table (compilerprocs) with all symbols that are marked as compilerproc. compilerprocs are needed by the code generator. A magic proc is not the same as a compilerproc: A magic is a proc that needs compiler magic for its semantic checking, a compilerproc is a proc that is used by the code generator.

Code generation for closures

Code generation for closures is implemented by lambda lifting.

Design

A closure proc var can call ordinary procs of the default Nim calling convention. But not the other way round! A closure is implemented as a tuple[prc, env]. env can be nil implying a call without a closure. This means that a call through a closure generates an if but the interoperability is worth the cost of the if. Thunk generation would be possible too, but it's slightly more effort to implement.

Tests with GCC on Amd64 showed that it's really beneficial if the 'environment' pointer is passed as the last argument, not as the first argument.

Proper thunk generation is harder because the proc that is to wrap could stem from a complex expression:

receivesClosure(returnsDefaultCC[i])

A thunk would need to call returnsDefaultCC[i] somehow and that would require an additional closure generation... Ok, not really, but it requires to pass the function to call. So we'd end up with 2 indirect calls instead of one. Another much more severe problem with this solution is that it's not GC-safe to pass a proc pointer around via a generic ref type.

Example code:

proc add(x: int): proc (y: int): int {.closure.} =
  return proc (y: int): int =
    return x + y

var add2 = add(2)
echo add2(5) #OUT 7

This should produce roughly this code:

type
  Env = ref object
    x: int # data

proc anon(y: int, c: Env): int =
  return y + c.x

proc add(x: int): tuple[prc, data] =
  var env: Env
  new env
  env.x = x
  result = (anon, env)

var add2 = add(2)
let tmp = if add2.data == nil: add2.prc(5) else: add2.prc(5, add2.data)
echo tmp

Beware of nesting:

proc add(x: int): proc (y: int): proc (z: int): int {.closure.} {.closure.} =
  return lambda (y: int): proc (z: int): int {.closure.} =
    return lambda (z: int): int =
      return x + y + z

var add24 = add(2)(4)
echo add24(5) #OUT 11

This should produce roughly this code:

type
  EnvX = ref object
    x: int # data
  
  EnvY = ref object
    y: int
    ex: EnvX

proc lambdaZ(z: int, ey: EnvY): int =
  return ey.ex.x + ey.y + z

proc lambdaY(y: int, ex: EnvX): tuple[prc, data: EnvY] =
  var ey: EnvY
  new ey
  ey.y = y
  ey.ex = ex
  result = (lambdaZ, ey)

proc add(x: int): tuple[prc, data: EnvX] =
  var ex: EnvX
  ex.x = x
  result = (lambdaY, ex)

var tmp = add(2)
var tmp2 = tmp.fn(4, tmp.data)
var add24 = tmp2.fn(4, tmp2.data)
echo add24(5)

We could get rid of nesting environments by always inlining inner anon procs. More useful is escape analysis and stack allocation of the environment, however.

Accumulator

proc getAccumulator(start: int): proc (): int {.closure} =
  var i = start
  return lambda: int =
    inc i
    return i

proc p =
  var delta = 7
  proc accumulator(start: int): proc(): int =
    var x = start-1
    result = proc (): int =
      x = x + delta
      inc delta
      return x
  
  var a = accumulator(3)
  var b = accumulator(4)
  echo a() + b()

Internals

Lambda lifting is implemented as part of the transf pass. The transf pass generates code to set up the environment and to pass it around. However, this pass does not change the types! So we have some kind of mismatch here; on the one hand the proc expression becomes an explicit tuple, on the other hand the tyProc(ccClosure) type is not changed. For C code generation it's also important the hidden formal param is void* and not something more specialized. However, the more specialized env type needs to passed to the backend somehow. We deal with this by modifying s.ast[paramPos] to contain the formal hidden parameter, but not s.typ!

Notes on type and AST representation

To be expanded.

Integer literals

In Nim, there is a redundant way to specify the type of an integer literal. First, it should be unsurprising that every node has a node kind. The node of an integer literal can be any of the following values:

nkIntLit, nkInt8Lit, nkInt16Lit, nkInt32Lit, nkInt64Lit,
nkUIntLit, nkUInt8Lit, nkUInt16Lit, nkUInt32Lit, nkUInt64Lit

On top of that, there is also the typ field for the type. The kind of the typ field can be one of the following ones, and it should be matching the literal kind:

tyInt, tyInt8, tyInt16, tyInt32, tyInt64, tyUInt, tyUInt8,
tyUInt16, tyUInt32, tyUInt64

Then there is also the integer literal type. This is a specific type that is implicitly convertible into the requested type if the requested type can hold the value. For this to work, the type needs to know the concrete value of the literal. For example an expression 321 will be of type int literal(321). This type is implicitly convertible to all integer types and ranges that contain the value 321. That would be all builtin integer types except uint8 and int8 where 321 would be out of range. When this literal type is assigned to a new var or let variable, it's type will be resolved to just int, not int literal(321) unlike constants. A constant keeps the full int literal(321) type. Here is an example where that difference matters.

proc foo(arg: int8) =
  echo "def"

const tmp1 = 123
foo(tmp1)  # OK

let tmp2 = 123
foo(tmp2) # Error

In a context with multiple overloads, the integer literal kind will always prefer the int type over all other types. If none of the overloads is of type int, then there will be an error because of ambiguity.

proc foo(arg: int) =
  echo "abc"
proc foo(arg: int8) =
  echo "def"
foo(123) # output: abc

proc bar(arg: int16) =
  echo "abc"
proc bar(arg: int8) =
  echo "def"

bar(123) # Error ambiguous call

In the compiler these integer literal types are represented with the node kind nkIntLit, type kind tyInt and the member n of the type pointing back to the integer literal node in the ast containing the integer value. These are the properties that hold true for integer literal types.

n.kind == nkIntLit
n.typ.kind == tyInt
n.typ.n == n

Other literal types, such as uint literal(123) that would automatically convert to other integer types, but prefers to become a uint are not part of the Nim language.

In an unchecked AST, the typ field is nil. The type checker will set the typ field accordingly to the node kind. Nodes of kind nkIntLit will get the integer literal type (e.g. int literal(123)). Nodes of kind nkUIntLit will get type uint (kind tyUint), etc.

This also means that it is not possible to write a literal in an unchecked AST that will after sem checking just be of type int and not implicitly convertible to other integer types. This only works for all integer types that are not int.