5.1. Using GHC¶
5.1.1. Getting started: compiling programs¶
In this chapter you’ll find a complete reference to the GHC command-line syntax, including all 400+ flags. It’s a large and complex system, and there are lots of details, so it can be quite hard to figure out how to get started. With that in mind, this introductory section provides a quick introduction to the basic usage of GHC for compiling a Haskell program, before the following sections dive into the full syntax.
Let’s create a Hello World program, and compile and run it. First,
create a file hello.hs
containing the Haskell code:
main = putStrLn "Hello, World!"
To compile the program, use GHC like this:
$ ghc hello.hs
(where $
represents the prompt: don’t type it). GHC will compile the
source file hello.hs
, producing an object file hello.o
and an
interface file hello.hi
, and then it will link the object file to
the libraries that come with GHC to produce an executable called
hello
on Unix/Linux/Mac, or hello.exe
on Windows.
By default GHC will be very quiet about what it is doing, only printing
error messages. If you want to see in more detail what’s going on behind
the scenes, add -v
to the command line.
Then we can run the program like this:
$ ./hello
Hello World!
If your program contains multiple modules, then you only need to tell
GHC the name of the source file containing the Main
module, and GHC
will examine the import
declarations to find the other modules that
make up the program and find their source files. This means that, with
the exception of the Main
module, every source file should be named
after the module name that it contains (with dots replaced by directory
separators). For example, the module Data.Person
would be in the
file Data/Person.hs
on Unix/Linux/Mac, or Data\Person.hs
on
Windows.
5.1.2. Options overview¶
GHC’s behaviour is controlled by options, which for historical reasons are also sometimes referred to as command-line flags or arguments. Options can be specified in three ways:
5.1.2.1. Command-line arguments¶
An invocation of GHC takes the following form:
ghc [argument...]
Command-line arguments are either options or file names.
Command-line options begin with -
. They may not be grouped:
-vO
is different from -v -O
. Options need not precede filenames:
e.g., ghc *.o -o foo
. All options are processed and then applied to
all files; you cannot, for example, invoke
ghc -c -O1 Foo.hs -O2 Bar.hs
to apply different optimisation levels
to the files Foo.hs
and Bar.hs
.
Note
Note that command-line options are order-dependent, with arguments being
evaluated from left-to-right. This can have seemingly strange effects in the
presence of flag implication. For instance, consider
-fno-specialise
and -O1
(which implies
-fspecialise
). These two command lines mean very different
things:
-fno-specialise -O1
-fspecialise
will be enabled as the-fno-specialise
is overridden by the-O1
.
-O1 -fno-specialise
-fspecialise
will not be enabled, since the-fno-specialise
overrides the-fspecialise
implied by-O1
.
5.1.2.2. Command line options in source files¶
Sometimes it is useful to make the connection between a source file and
the command-line options it requires quite tight. For instance, if a
Haskell source file deliberately uses name shadowing, it should be
compiled with the -Wno-name-shadowing
option. Rather than
maintaining the list of per-file options in a Makefile
, it is
possible to do this directly in the source file using the
OPTIONS_GHC
pragma
{-# OPTIONS_GHC -Wno-name-shadowing #-}
module X where
...
OPTIONS_GHC
is a file-header pragma (see OPTIONS_GHC pragma).
Only dynamic flags can be used in an OPTIONS_GHC
pragma (see
Dynamic and Mode options).
Note that your command shell does not get to the source file options,
they are just included literally in the array of command-line arguments
the compiler maintains internally, so you’ll be desperately disappointed
if you try to glob etc. inside OPTIONS_GHC
.
Note
The contents of OPTIONS_GHC
are appended to the command-line
options, so options given in the source file override those given on the
command-line.
It is not recommended to move all the contents of your Makefiles into
your source files, but in some circumstances, the OPTIONS_GHC
pragma
is the Right Thing. (If you use -keep-hc-file
and have OPTION
flags in
your module, the OPTIONS_GHC
will get put into the generated .hc
file).
5.1.3. Dynamic and Mode options¶
Each of GHC’s command line options is classified as dynamic or mode:
Mode: A mode may be used on the command line only. You can pass only one mode flag. For example,
--make
or-E
. The available modes are listed in Modes of operation.Dynamic: A dynamic flag may be used on the command line, in a
OPTIONS_GHC
pragma in a source file, or set using:set
in GHCi.
The flag reference tables (Flag reference) lists the status of each flag.
5.1.4. Meaningful file suffixes¶
File names with “meaningful” suffixes (e.g., .lhs
or .o
) cause
the “right thing” to happen to those files.
.hs
- A Haskell module.
.lhs
A “literate Haskell” module.
.hspp
- A file created by the preprocessor.
.hi
- A Haskell interface file, probably compiler-generated.
.hie
- An extended Haskell interface file, produced by the Haskell compiler.
.hc
- Intermediate C file produced by the Haskell compiler.
.c
- A C file not produced by the Haskell compiler.
.ll
- An llvm-intermediate-language source file, usually produced by the compiler.
.bc
- An llvm-intermediate-language bitcode file, usually produced by the compiler.
.s
- An assembly-language source file, usually produced by the compiler.
.o
- An object file, produced by an assembler.
Files with other suffixes (or without suffixes) are passed straight to the linker.
5.1.5. Modes of operation¶
GHC’s behaviour is firstly controlled by a mode flag. Only one of these flags may be given, but it does not necessarily need to be the first option on the command-line. For instance,
$ ghc Main.hs --make -o my-application
If no mode flag is present, then GHC will enter --make
mode
(Using ghc –make) if there are any Haskell source files given on the
command line, or else it will link the objects named on the command line
to produce an executable.
The available mode flags are:
-
--interactive
¶ Interactive mode, which is also available as ghci. Interactive mode is described in more detail in Using GHCi.
-
--make
¶ In this mode, GHC will build a multi-module Haskell program automatically, figuring out dependencies for itself. If you have a straightforward Haskell program, this is likely to be much easier, and faster, than using make. Make mode is described in Using ghc –make.
This mode is the default if there are any Haskell source files mentioned on the command line, and in this case the
--make
option can be omitted.
-
-e
⟨expr⟩
¶ Expression-evaluation mode. This is very similar to interactive mode, except that there is a single expression to evaluate (⟨expr⟩) which is given on the command line. See Expression evaluation mode for more details.
-
-E
¶ Stop after preprocessing (
.hspp
file)
-
-C
¶ Stop after generating C (
.hc
file)
-
-S
¶ Stop after generating assembly (
.s
file)
-
-c
¶ Stop after generating object (
.o
) fileThis is the traditional batch-compiler mode, in which GHC can compile source files one at a time, or link objects together into an executable. See Batch compiler mode.
-
-M
¶ Dependency-generation mode. In this mode, GHC can be used to generate dependency information suitable for use in a
Makefile
. See Dependency generation.
-
--frontend
⟨module⟩
¶ Run GHC using the given frontend plugin. See Frontend plugins for details.
-
--mk-dll
¶ DLL-creation mode (Windows only). See Creating a DLL.
-
--show-iface
⟨file⟩
¶ Read the interface in ⟨file⟩ and dump it as text to
stdout
. For exampleghc --show-iface M.hi
.
-
--show-options
¶ Print the supported command line options. This flag can be used for autocompletion in a shell.
-
--info
¶ Print information about the compiler.
-
--numeric-version
¶ Print GHC’s numeric version number only.
-
--print-libdir
¶ Print the path to GHC’s library directory. This is the top of the directory tree containing GHC’s libraries, interfaces, and include files (usually something like
/usr/local/lib/ghc-5.04
on Unix). This is the value of$libdir
in the package configuration file (see Packages).
5.1.5.1. Using ghc
--make
¶
In this mode, GHC will build a multi-module Haskell program by following
dependencies from one or more root modules (usually just Main
). For
example, if your Main
module is in a file called Main.hs
, you
could compile and link the program like this:
ghc --make Main.hs
In fact, GHC enters make mode automatically if there are any Haskell source files on the command line and no other mode is specified, so in this case we could just type
ghc Main.hs
Any number of source file names or module names may be specified; GHC
will figure out all the modules in the program by following the imports
from these initial modules. It will then attempt to compile each module
which is out of date, and finally, if there is a Main
module, the
program will also be linked into an executable.
The main advantages to using ghc --make
over traditional
Makefile
s are:
GHC doesn’t have to be restarted for each compilation, which means it can cache information between compilations. Compiling a multi-module program with
ghc --make
can be up to twice as fast as runningghc
individually on each source file.You don’t have to write a
Makefile
.GHC re-calculates the dependencies each time it is invoked, so the dependencies never get out of sync with the source.
Using the
-j[⟨n⟩]
flag, you can compile modules in parallel. Specify-j ⟨n⟩
to compile ⟨n⟩ jobs in parallel. If ⟨n⟩ is omitted, then it defaults to the number of processors.
Any of the command-line options described in the rest of this chapter
can be used with --make
, but note that any options you give on the
command line will apply to all the source files compiled, so if you want
any options to apply to a single source file only, you’ll need to use an
OPTIONS_GHC
pragma (see Command line options in source files).
If the program needs to be linked with additional objects (say, some auxiliary C code), then the object files can be given on the command line and GHC will include them when linking the executable.
For backward compatibility with existing make scripts, when used in
combination with -c
, the linking phase is omitted (same as
--make -no-link
).
Note that GHC can only follow dependencies if it has the source file available, so if your program includes a module for which there is no source file, even if you have an object and an interface file for the module, then GHC will complain. The exception to this rule is for package modules, which may or may not have source files.
The source files for the program don’t all need to be in the same
directory; the -i
option can be used to add directories to the
search path (see The search path).
-
-j
[⟨n⟩]
¶ Perform compilation in parallel when possible. GHC will use up to ⟨N⟩ threads during compilation. If N is omitted, then it defaults to the number of processors. Note that compilation of a module may not begin until its dependencies have been built.
5.1.5.2. Expression evaluation mode¶
This mode is very similar to interactive mode, except that there is a
single expression to evaluate which is specified on the command line as
an argument to the -e
option:
ghc -e expr
Haskell source files may be named on the command line, and they will be loaded exactly as in interactive mode. The expression is evaluated in the context of the loaded modules.
For example, to load and run a Haskell program containing a module
Main
, we might say:
ghc -e Main.main Main.hs
or we can just use this mode to evaluate expressions in the context of
the Prelude
:
$ ghc -e "interact (unlines.map reverse.lines)"
hello
olleh
5.1.5.3. Batch compiler mode¶
In batch mode, GHC will compile one or more source files given on the command line.
The first phase to run is determined by each input-file suffix, and the last phase is determined by a flag. If no relevant flag is present, then go all the way through to linking. This table summarises:
Phase of the compilation system | Suffix saying “start here” | Flag saying “stop after” | (suffix of) output file |
---|---|---|---|
literate pre-processor | .lhs |
.hs |
|
C pre-processor (opt.) | .hs (with -cpp ) |
-E |
.hspp |
Haskell compiler | .hs |
-C , -S |
.hc , .s |
C compiler (opt.) | .hc or .c |
-S |
.s |
assembler | .s |
-c |
.o |
linker | ⟨other⟩ | a.out |
Thus, a common invocation would be:
ghc -c Foo.hs
to compile the Haskell source file Foo.hs
to an object file
Foo.o
.
Note
What the Haskell compiler proper produces depends on what backend code generator is used. See GHC Backends for more details.
Note
Pre-processing is optional, the -cpp
flag turns it
on. See Options affecting the C pre-processor for more details.
Note
The option -E
runs just the pre-processing passes of
the compiler, dumping the result in a file.
Note
The option -C
is only available when GHC is built in
unregisterised mode. See Unregisterised compilation for more details.
5.1.5.3.1. Overriding the default behaviour for a file¶
As described above, the way in which a file is processed by GHC depends on its
suffix. This behaviour can be overridden using the -x ⟨suffix⟩
option:
-
-x
⟨suffix⟩
¶ Causes all files following this option on the command line to be processed as if they had the suffix ⟨suffix⟩. For example, to compile a Haskell module in the file
M.my-hs
, useghc -c -x hs M.my-hs
.
5.1.6. Verbosity options¶
See also the --help
, --version
, --numeric-version
, and
--print-libdir
modes in Modes of operation.
-
-v
¶ The
-v
option makes GHC verbose: it reports its version number and shows (on stderr) exactly how it invokes each phase of the compilation system. Moreover, it passes the-v
flag to most phases; each reports its version number (and possibly some other information).Please, oh please, use the
-v
option when reporting bugs! Knowing that you ran the right bits in the right order is always the first thing we want to verify.
-
-v⟨n⟩
¶ To provide more control over the compiler’s verbosity, the
-v
flag takes an optional numeric argument. Specifying-v
on its own is equivalent to-v3
, and the other levels have the following meanings:-v0
- Disable all non-essential messages (this is the default).
-v1
- Minimal verbosity: print one line per compilation (this is the
default when
--make
or--interactive
is on). -v2
- Print the name of each compilation phase as it is executed.
(equivalent to
-dshow-passes
). -v3
- The same as
-v2
, except that in addition the full command line (if appropriate) for each compilation phase is also printed. -v4
- The same as
-v3
except that the intermediate program representation after each compilation phase is also printed (excluding preprocessed and C/assembly files).
-
-fprint-potential-instances
¶ When GHC can’t find an instance for a class, it displays a short list of some in the instances it knows about. With this flag it prints all the instances it knows about.
-
-fhide-source-paths
¶ Starting with minimal verbosity (
-v1
, see-v
), GHC displays the name, the source path and the target path of each compiled module. This flag can be used to reduce GHC’s output by hiding source paths and target paths.
The following flags control the way in which GHC displays types in error messages and in GHCi:
-
-fprint-unicode-syntax
¶ When enabled GHC prints type signatures using the unicode symbols from the
UnicodeSyntax
extension. For instance,ghci> :set -fprint-unicode-syntax ghci> :t +v (>>) (>>) ∷ Monad m ⇒ ∀ a b. m a → m b → m b
-
-fprint-explicit-foralls
¶ Using
-fprint-explicit-foralls
makes GHC print explicitforall
quantification at the top level of a type; normally this is suppressed. For example, in GHCi:ghci> let f x = x ghci> :t f f :: a -> a ghci> :set -fprint-explicit-foralls ghci> :t f f :: forall a. a -> a
However, regardless of the flag setting, the quantifiers are printed under these circumstances:
For nested
foralls
, e.g.ghci> :t GHC.ST.runST GHC.ST.runST :: (forall s. GHC.ST.ST s a) -> a
If any of the quantified type variables has a kind that mentions a kind variable, e.g.
ghci> :i Data.Type.Equality.sym Data.Type.Equality.sym :: forall k (a :: k) (b :: k). (a Data.Type.Equality.:~: b) -> b Data.Type.Equality.:~: a -- Defined in Data.Type.Equality
-
-fprint-explicit-kinds
¶ Using
-fprint-explicit-kinds
makes GHC print kind arguments in types, which are normally suppressed. This can be important when you are using kind polymorphism. For example:ghci> :set -XPolyKinds ghci> data T a (b :: l) = MkT ghci> :t MkT MkT :: forall k l (a :: k) (b :: l). T a b ghci> :set -fprint-explicit-kinds ghci> :t MkT MkT :: forall k l (a :: k) (b :: l). T @{k} @l a b ghci> :set -XNoPolyKinds ghci> :t MkT MkT :: T @{*} @* a b
In the output above, observe that
T
has two kind variables (k
andl
) and two type variables (a
andb
). Note thatk
is an inferred variable andl
is a specified variable (see Inferred vs. specified type variables), so as a result, they are displayed using slightly different syntax in the typeT @{k} @l a b
. The application ofl
(with@l
) is the standard syntax for visible type application (see Visible type application). The application ofk
(with@{k}
), however, uses a hypothetical syntax for visible type application of inferred type variables. This syntax is not currently exposed to the programmer, but it is nevertheless displayed when-fprint-explicit-kinds
is enabled.
-
-fprint-explicit-coercions
¶ Using
-fprint-explicit-coercions
makes GHC print coercions in types. When trying to prove the equality between types of different kinds, GHC uses type-level coercions. Users will rarely need to see these, as they are meant to be internal.
-
-fprint-axiom-incomps
¶ Using
-fprint-axiom-incomps
tells GHC to display incompatibilities between closed type families’ equations, whenever they are printed by:info
or--show-iface ⟨file⟩
.ghci> :i Data.Type.Equality.== type family (==) (a :: k) (b :: k) :: Bool where (==) (f a) (g b) = (f == g) && (a == b) (==) a a = 'True (==) _1 _2 = 'False ghci> :set -fprint-axiom-incomps ghci> :i Data.Type.Equality.== type family (==) (a :: k) (b :: k) :: Bool where {- #0 -} (==) (f a) (g b) = (f == g) && (a == b) {- #1 -} (==) a a = 'True -- incompatible with: #0 {- #2 -} (==) _1 _2 = 'False -- incompatible with: #1, #0
The equations are numbered starting from 0, and the comment after each equation refers to all preceding equations it is incompatible with.
-
-fprint-equality-relations
¶ Using
-fprint-equality-relations
tells GHC to distinguish between its equality relations when printing. For example,~
is homogeneous lifted equality (the kinds of its arguments are the same) while~~
is heterogeneous lifted equality (the kinds of its arguments might be different) and~#
is heterogeneous unlifted equality, the internal equality relation used in GHC’s solver. Generally, users should not need to worry about the subtleties here;~
is probably what you want. Without-fprint-equality-relations
, GHC prints all of these as~
. See also Equality constraints.
-
-fprint-expanded-synonyms
¶ When enabled, GHC also prints type-synonym-expanded types in type errors. For example, with this type synonyms:
type Foo = Int type Bar = Bool type MyBarST s = ST s Bar
This error message:
Couldn't match type 'Int' with 'Bool' Expected type: ST s Foo Actual type: MyBarST s
Becomes this:
Couldn't match type 'Int' with 'Bool' Expected type: ST s Foo Actual type: MyBarST s Type synonyms expanded: Expected type: ST s Int Actual type: ST s Bool
-
-fprint-typechecker-elaboration
¶ When enabled, GHC also prints extra information from the typechecker in warnings. For example:
main :: IO () main = do return $ let a = "hello" in a return ()
This warning message:
A do-notation statement discarded a result of type ‘[Char]’ Suppress this warning by saying ‘_ <- ($) return let a = "hello" in a’ or by using the flag -fno-warn-unused-do-bind
Becomes this:
A do-notation statement discarded a result of type ‘[Char]’ Suppress this warning by saying ‘_ <- ($) return let AbsBinds [] [] {Exports: [a <= a <>] Exported types: a :: [Char] [LclId, Str=DmdType] Binds: a = "hello"} in a’ or by using the flag -fno-warn-unused-do-bind
-
-fdefer-diagnostics
¶ Causes GHC to group diagnostic messages by severity and output them after other messages when building a multi-module Haskell program. This flag can make diagnostic messages more visible when used in conjunction with
--make
and-j[⟨n⟩]
. Otherwise, it can be hard to find the relevant errors or likely to ignore the warnings when they are mixed with many other messages.
-
-fdiagnostics-color
=⟨always|auto|never⟩
¶ Causes GHC to display error messages with colors. To do this, the terminal must have support for ANSI color codes, or else garbled text will appear. The default value is
auto
, which means GHC will make an attempt to detect whether terminal supports colors and choose accordingly.The precise color scheme is controlled by the environment variable
GHC_COLORS
(orGHC_COLOURS
). This can be set to colon-separated list ofkey=value
pairs. These are the default settings:header=:message=1:warning=1;35:error=1;31:fatal=1;31:margin=1;34
Each value is expected to be a Select Graphic Rendition (SGR) substring. The formatting of each element can inherit from parent elements. For example, if
header
is left empty, it will inherit the formatting ofmessage
. Alternatively ifheader
is set to1
(bold), it will be bolded but still inherits the color ofmessage
.Currently, in the primary message, the following inheritance tree is in place:
message
header
warning
error
fatal
In the caret diagnostics, there is currently no inheritance at all between
margin
,warning
,error
, andfatal
.The environment variable can also be set to the magical values
never
oralways
, which is equivalent to setting the corresponding-fdiagnostics-color
flag but with lower precedence.
-
-fdiagnostics-show-caret
¶ Controls whether GHC displays a line of the original source code where the error was detected. This also affects the associated caret symbol that points at the region of code at fault. The flag is on by default.
-
-ferror-spans
¶ Causes GHC to emit the full source span of the syntactic entity relating to an error message. Normally, GHC emits the source location of the start of the syntactic entity only.
For example:
test.hs:3:6: parse error on input `where'
becomes:
test296.hs:3:6-10: parse error on input `where'
And multi-line spans are possible too:
test.hs:(5,4)-(6,7): Conflicting definitions for `a' Bound at: test.hs:5:4 test.hs:6:7 In the binding group for: a, b, a
Note that line numbers start counting at one, but column numbers start at zero. This choice was made to follow existing convention (i.e. this is how Emacs does it).
-
-fkeep-going
¶ Since: 8.10.1 Causes GHC to continue the compilation if a module has an error. Any reverse dependencies are pruned immediately and the whole compilation is still flagged as an error. This option has no effect if parallel compilation (
-j[⟨n⟩]
) is in use.
-
-freverse-errors
¶ Causes GHC to output errors in reverse line-number order, so that the errors and warnings that originate later in the file are displayed first.
-
-H
⟨size⟩
¶ Set the minimum size of the heap to ⟨size⟩. This option is equivalent to
+RTS -Hsize
, see RTS options to control the garbage collector.
-
-Rghc-timing
¶ Prints a one-line summary of timing statistics for the GHC run. This option is equivalent to
+RTS -tstderr
, see RTS options to control the garbage collector.
5.1.7. Platform-specific Flags¶
Some flags only make sense for particular target platforms.
-
-msse2
¶ (x86 only, added in GHC 7.0.1) Use the SSE2 registers and instruction set to implement floating point operations when using the native code generator. This gives a substantial performance improvement for floating point, but the resulting compiled code will only run on processors that support SSE2 (Intel Pentium 4 and later, or AMD Athlon 64 and later). The LLVM backend will also use SSE2 if your processor supports it but detects this automatically so no flag is required.
SSE2 is unconditionally used on x86-64 platforms.
-
-msse4.2
¶ (x86 only, added in GHC 7.4.1) Use the SSE4.2 instruction set to implement some floating point and bit operations when using the native code generator. The resulting compiled code will only run on processors that support SSE4.2 (Intel Core i7 and later). The LLVM backend will also use SSE4.2 if your processor supports it but detects this automatically so no flag is required.
-
-mbmi2
¶ (x86 only, added in GHC 7.4.1) Use the BMI2 instruction set to implement some bit operations when using the native code generator. The resulting compiled code will only run on processors that support BMI2 (Intel Haswell and newer, AMD Excavator, Zen and newer).
5.1.8. Haddock¶
-
-haddock
¶ By default, GHC ignores Haddock comments (
-- | ...
and-- ^ ...
) and does not check that they’re associated with a valid term, such as a top-level type-signature. With this flag GHC will parse Haddock comments and include them in the interface file it produces.Note that this flag makes GHC’s parser more strict so programs which are accepted without Haddock may be rejected with
-haddock
.
5.1.9. Miscellaneous flags¶
Some flags only make sense for a particular use case.
-
-ghcversion-file
⟨path to ghcversion.h⟩
¶ When GHC is used to compile C files, GHC adds package include paths and includes
ghcversion.h
directly. The compiler will lookup the path for theghcversion.h
file from therts
package in the package database. In some cases, the compiler’s package database does not contain therts
package, or one wants to specify a specificghcversions.h
to be included. This option can be used to specify the path to theghcversions.h
file to be included. This is primarily intended to be used by GHC’s build system.
5.1.9.1. Other environment variables¶
GHC can also be configured using environment variables. Currently the only
variable it supports is GHC_NO_UNICODE
, which, when set, disables Unicode
output regardless of locale settings. GHC_NO_UNICODE
can be set to anything
+(event an empty string) to trigger this behaviour.