User guide

Stack is a modern, cross-platform build tool for Haskell code.

This guide takes a new Stack user through the typical workflows. This guide will not teach Haskell or involve much code, and it requires no prior experience with the Haskell packaging system or other build tools. Terms used in the guide are defined in the glossary.

Stack's functions

Stack handles the management of your toolchain (including GHC — the Glasgow Haskell Compiler — and, for Windows users, MSYS2), building and registering libraries, building build tool dependencies, and more. While it can use existing tools on your system, Stack has the capacity to be your one-stop shop for all Haskell tooling you need. This guide will follow that Stack-centric approach.

What makes Stack special?

The primary Stack design point is reproducible builds. If you run stack build today, you should get the same result running stack build tomorrow. There are some cases that can break that rule (changes in your operating system configuration, for example), but, overall, Stack follows this design philosophy closely. To make this a simple process, Stack uses curated package sets called snapshots.

Stack has also been designed from the ground up to be user friendly, with an intuitive, discoverable command line interface. For many users, simply downloading Stack and reading stack --help will be enough to get up and running. This guide provides a more gradual tour for users who prefer that learning style.

To build your project, Stack uses a project-level configuration file, named stack.yaml, in the root directory of your project as a sort of blueprint. That file contains a reference, called a resolver, to the snapshot which your package will be built against.

Finally, Stack is isolated: it will not make changes outside of specific Stack directories. Stack-built files generally go in either the Stack root directory (default: ~/.stack on Unix-like operating systems, or, %LOCALAPPDATA%\Programs\stack on Windows) or ./.stack-work directories local to each project. The Stack root directory holds packages belonging to snapshots and any Stack-installed versions of GHC. Stack will not tamper with any system version of GHC or interfere with packages installed by other build tools (such as Cabal (the tool)).

Downloading and Installation

The documentation dedicated to downloading Stack has the most up-to-date information for a variety of operating systems. Instead of repeating that content here, please go check out that page and come back here when you can successfully run stack --version.

We also assume that the directory reported by stack path --local-bin has been added to the PATH.

Hello World Example

With Stack installed, let's create a new project from a template and walk through the most common Stack commands.

In this guide, an initial $ represents the command line prompt. The prompt may differ in the terminal on your operating system. Unless stated otherwise, the working directory is the project's root directory.

The stack new command

We'll start off with the stack new command to create a new project, that will contain a Haskell package of the same name. So let's pick a valid package name first:

A package is identified by a globally-unique package name, which consists of one or more alphanumeric words separated by hyphens. To avoid ambiguity, each of these words should contain at least one letter.

(From the Cabal users guide)

We'll call our project helloworld, and we'll use the new-template project template. This template is used by default, but in our example we will refer to it expressly. From the root directory for all our Haskell projects, we command:

$ stack new helloworld new-template

For this first Stack command, there's quite a bit of initial setup it needs to do (such as downloading the list of packages available upstream), so you'll see a lot of output. Over the course of this guide a lot of the content will begin to make more sense.

We now have a project in the helloworld directory! We will change to that directory, with:

$ cd helloworld

The stack build command

Next, we'll run the most important Stack command: stack build.

$ stack build
# installing ... building ...

Stack needs a version of GHC in order to build your project. Stack will discover that you are missing it and will install it for you. You can do this manually by using the stack setup command.

You'll get intermediate download percentage statistics while the download is occurring. This command may take some time, depending on download speeds.

NOTE: GHC will be installed to your Stack programs directory, so calling ghc on the command line won't work. See the stack exec, stack ghc, and stack runghc commands below for more information.

Once a version of GHC is installed, Stack will then build your project.

The stack exec command

Looking closely at the output of the previous command, you can see that it built both a library called helloworld and an executable called helloworld-exe (on Windows, helloworld-exe.exe). We'll explain more in the next section, but, for now, just notice that the executables are installed in a location in our project's .stack-work directory.

Now, Let's use stack exec to run our executable (which just outputs "someFunc"):

$ stack exec helloworld-exe
someFunc

stack exec works by providing the same reproducible environment that was used to build your project to the command that you are running. Thus, it knew where to find helloworld-exe even though it is hidden in the .stack-work directory.

The stack test command

Finally, like all good software, helloworld actually has a test suite.

Let's run it with stack test:

$ stack test
# build output ...

Reading the output, you'll see that Stack first builds the test suite and then automatically runs it for us. For both the build and test command, already built components are not built again. You can see this by running stack build and stack test a second time:

$ stack build
$ stack test
# build output ...

Inner Workings of Stack

In this subsection, we'll dissect the helloworld example in more detail.

Files in helloworld

Before studying Stack more, let's understand our project a bit better. The files in the directory include:

app/Main.hs
src/Lib.hs
test/Spec.hs
ChangeLog.md
README.md
LICENSE
Setup.hs
helloworld.cabal
package.yaml
stack.yaml
.gitignore

The app/Main.hs, src/Lib.hs, and test/Spec.hs files are all Haskell source files that compose the actual functionality of our project (we won't dwell on them here).

The ChangeLog.md, README.md, LICENSE and .gitignore files have no effect on the build.

The helloworld.cabal file is updated automatically as part of the stack build process and should not be modified.

The files of interest here are Setup.hs, stack.yaml, and package.yaml.

The Setup.hs file is a component of the Cabal build system which Stack uses. It's technically not needed by Stack, but it is still considered good practice in the Haskell world to include it. The file we're using is straight boilerplate:

import Distribution.Simple
main = defaultMain

Next, let's look at our stack.yaml file, which gives our project-level settings. Ignoring comments beginning #, the contents will look something like this:

resolver:
  url: https://raw.githubusercontent.com/commercialhaskell/stackage-snapshots/master/lts/19/17.yaml
packages:
- .

The value of the resolver key tells Stack how to build your package: which GHC version to use, versions of package dependencies, and so on. Our value here says to use LTS Haskell 19.7, which implies GHC 9.0.2 (which is why stack setup installs that version of GHC). There are a number of values you can use for resolver, which we'll cover later.

The value of the packages key tells Stack which local packages to build. In our simple example, we have only a single package in our project, located in the same directory, so '.' suffices. However, Stack has powerful support for multi-package projects, which we'll elaborate on as this guide progresses.

Another file important to the build is package.yaml.

The package.yaml file describes the package in the Hpack format. Stack has in-built Hpack functionality and this is its preferred package format. The default behaviour is to generate the Cabal file (here named helloworld.cabal) from this package.yaml file, and accordingly you should not modify the Cabal file.

It is also important to remember that Stack is built on top of the Cabal build system. Therefore, an understanding of the moving parts in Cabal are necessary. In Cabal, we have individual packages, each of which contains a single Cabal file, named <package_name>.cabal. The Cabal file can define one or more components: a library, executables, test suites, and benchmarks. It also specifies additional information such as library dependencies, default language pragmas, and so on.

In this guide, we'll discuss the bare minimum necessary to understand how to modify a package.yaml file. You can see a full list of the available options at the Hpack documentation. The Cabal User Guide the definitive reference for the Cabal file format.

The stack setup command

As we saw above, the build command installed GHC for us. Just for kicks, let's manually run the setup command:

$ stack setup
stack will use a sandboxed GHC it installed
For more information on paths, see 'stack path' and 'stack exec env'
To use this GHC and packages outside of a project, consider using:
stack ghc, stack ghci, stack runghc, or stack exec

Thankfully, the command is smart enough to know not to perform an installation twice. As the command output above indicates, you can use stack path for quite a bit of path information (which we'll play with more later).

For now, we'll just look at where GHC is installed:

On Unix-like operating systems, command:

$ stack exec -- which ghc
/home/<user_name>/.stack/programs/x86_64-linux/ghc-9.0.2/bin/ghc

On Windows (with PowerShell), command:

$ stack exec -- where.exe ghc
C:\Users\<user_name>\AppData\Local\Programs\stack\x86_64-windows\ghc-9.0.2\bin\ghc.exe

As you can see from that path (and as emphasized earlier), the installation is placed to not interfere with any other GHC installation, whether system-wide or even different GHC versions installed by Stack.

Cleaning your project

You can clean up build artifacts for your project using the stack clean and stack purge commands.

The stack clean command

stack clean deletes the local working directories containing compiler output. By default, that means the contents of directories in .stack-work/dist, for all the .stack-work directories within a project.

Use stack clean <specific-package> to delete the output for the package specific-package only.

The stack purge command

stack purge deletes the local stack working directories, including extra-deps, git dependencies and the compiler output (including logs). It does not delete any snapshot packages, compilers or programs installed using stack install. This essentially reverts the project to a completely fresh state, as if it had never been built. stack purge is just a shortcut for stack clean --full

The stack build command

The build command is the heart and soul of Stack. It is the engine that powers building your code, testing it, getting dependencies, and more. Quite a bit of the remainder of this guide will cover more advanced build functions and features, such as building test and Haddocks at the same time, or constantly rebuilding blocking on file changes.

On a philosophical note: Running the build command twice with the same options and arguments should generally be a no-op (besides things like rerunning test suites), and should, in general, produce a reproducible result between different runs.

Adding dependencies

Let's say we decide to modify our helloworld source a bit to use a new library, perhaps the ubiquitous text package. In src/Lib.hs, we can, for example add:

{-# LANGUAGE OverloadedStrings #-}
module Lib
    ( someFunc
    ) where

import qualified Data.Text.IO as T

someFunc :: IO ()
someFunc = T.putStrLn "someFunc"

When we try to build this, things don't go as expected:

$ stack build
# build failure output (abridged for clarity) ...
src\Lib.hs:6:1: error:
    Could not load module ‘Data.Text.IO’
    It is a member of the hidden package ‘text-1.2.5.0’.
    Perhaps you need to add ‘text’ to the build-depends in your .cabal file.
    Use -v (or `:set -v` in ghci) to see a list of the files searched for.
  |
6 | import qualified Data.Text.IO as T
  | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This means that the package containing the module in question is not available. To tell Stack to use text, you need to add it to your package.yaml file — specifically in your dependencies section, like this:

dependencies:
- base >= 4.7 && < 5
- text # added here

Now if we rerun stack build, we should get a successful result:

$ stack build
# build output ...

This output means that the text package was downloaded, configured, built, and locally installed. Once that was done, we moved on to building our local package (helloworld). At no point did we need to ask Stack to build dependencies — it does so automatically.

Listing Dependencies

Let's have Stack add a few more dependencies to our project. First, we'll include two new packages in the dependencies section for our library in our package.yaml:

dependencies:
- base >= 4.7 && < 5
- text
- filepath
- containers

After adding these two dependencies, we can again run stack build to have them installed:

$ stack build
# build output ...

Finally, to find out which versions of these libraries Stack installed, we can ask Stack to ls dependencies:

$ stack ls dependencies
# dependency output ...

extra-deps

Let's try a more off-the-beaten-track package: the joke acme-missiles package. Our source code is simple:

module Lib
    ( someFunc
    ) where

import Acme.Missiles

someFunc :: IO ()
someFunc = launchMissiles

Again, we add this new dependency to the package.yaml file like this:

dependencies:
- base >= 4.7 && < 5
- text
- filepath
- containers
- acme-missiles # added

However, rerunning stack build shows us the following error message:

$ stack build
# build failure output ...

It says that it was unable to construct the build plan.

This brings us to the next major topic in using Stack.

Curated package sets

Remember above when stack new selected some LTS resolver for us? That defined our build plan and available packages. When we tried using the text package, it just worked, because it was part of the LTS package set.

We've specified the acme-missiles package in the package.yaml file (see above), but acme-missiles is not part of that LTS package set, so building failed.

To add acme-missles to the available packages, we'll use the extra-deps key in the stack.yaml file. That key defines extra packages, not present in the resolver, that will be needed as dependencies. You can add this like so:

extra-deps:
- acme-missiles-0.3 # not in the LTS resolver

Now stack build will succeed.

With that out of the way, let's dig a little bit more into these package sets, also known as snapshots. We mentioned the LTS resolvers, and you can get quite a bit of information about it at https://www.stackage.org/lts, including:

  • The appropriate resolver value (resolver: lts-19.17, as is currently the latest LTS)
  • The GHC version used
  • A full list of all packages available in this snapshot
  • The ability to perform a Hoogle search on the packages in this snapshot
  • A list of all modules in a snapshot, which can be useful when trying to determine which package to add to your package.yaml file.

You can also see a list of all available snapshots. You'll notice two flavors: LTS (for "Long Term Support") and Nightly. You can read more about them on the LTS Haskell Github page. If you're not sure which to use, start with LTS Haskell (which Stack will lean towards by default as well).

Resolvers and changing your compiler version

Let's explore package sets a bit further. Instead of lts-19.17, let's change our stack.yaml file to use the latest nightly. Right now, this is currently 2022-07-31 - please see the resolve from the link above to get the latest.

Then, Rerunning stack build will produce:

$ stack build
# Downloaded nightly-2020-07-31 build plan.
# build output ...

We can also change resolvers on the command line, which can be useful in a Continuous Integration (CI) setting, like on Travis. For example:

$ stack --resolver lts-18.28 build
# Downloaded lts-18.28 build plan.
# build output ...

When passed on the command line, you also get some additional "short-cut" versions of resolvers: --resolver nightly will use the newest Nightly resolver available, --resolver lts will use the newest LTS, and --resolver lts-19 will use the newest LTS in the 19.x series. The reason these are only available on the command line and not in your stack.yaml file is that using them:

  1. Will slow down your build (since Stack then needs to download information on the latest available LTS each time it builds)
  2. Produces unreliable results (since a build run today may proceed differently tomorrow because of changes outside of your control)

Changing GHC versions

Finally, let's try using an older LTS snapshot. We'll use the newest 18.x snapshot:

$ stack --resolver lts-18 build
# build output ...

This succeeds, automatically installing the necessary GHC along the way. So, we see that different LTS versions use different GHC versions and Stack can handle that.

Other resolver values

We've mentioned nightly-YYYY-MM-DD and lts-X.Y values for the resolver. There are actually other options available, and the list will grow over time. At the time of writing:

  • ghc-X.Y.Z, for requiring a specific GHC version but no additional packages
  • Experimental custom snapshot support

The most up-to-date information can always be found in the stack.yaml documentation.

Existing projects

Alright, enough playing around with simple projects. Let's take an open source package and try to build it. We'll be ambitious and use yackage, a local package server using Yesod. To get the code, we'll use the stack unpack command from the root directory for all our Haskell projects:

$ stack unpack yackage
Unpacked yackage-0.8.1 to <root_directory>/yackage-0.8.1/

You can also unpack to the directory of your liking instead of the current one by issuing:

$ stack unpack yackage --to <desired_directory>

This will create a yackage-0.8.1 directory inside <desired_directory>.

We will change to that directory, with:

$ cd yackage-0.8.1

The stack init command

This new directory does not have a stack.yaml file, so we need to make one first. We could do it by hand, but let's be lazy instead with the stack init command:

$ stack init
# init output ...

stack init does quite a few things for you behind the scenes:

  • Finds all of the Cabal files in your current directory and subdirectories (unless you use --ignore-subdirs) and determines the packages and versions they require
  • Finds the best combination of snapshot and package flags that allows everything to compile with minimum external dependencies
  • It tries to look for the best matching snapshot from latest LTS, latest nightly, other LTS versions in that order

Assuming it finds a match, it will write your stack.yaml file, and everything will work.

(Note: yackage does not currently support Hpack, but you can also use hpack-convert should you need to generate a package.yaml.)

Excluded Packages

Sometimes multiple packages in your project may have conflicting requirements. In that case stack init will fail, so what do you do?

You could manually create stack.yaml by omitting some packages to resolve the conflict. Alternatively you can ask stack init to do that for you by specifying --omit-packages flag on the command line. Let's see how that works.

To simulate a conflict we will use acme-missiles-0.3 in yackage and we will also copy yackage.cabal to another directory and change the name of the file and package to yackage-test. In this new package we will use acme-missiles-0.2 instead. Let's see what happens when we re-run stack init:

$ stack init --force --omit-packages
# init failure output ...

Looking at stack.yaml, you will see that the excluded packages have been commented out under the packages field. In case wrong packages are excluded you can uncomment the right one and comment the other one.

Packages may get excluded due to conflicting requirements among user packages or due to conflicting requirements between a user package and the resolver compiler. If all of the packages have a conflict with the compiler then all of them may get commented out.

When packages are commented out you will see a warning every time you run a command which needs the configuration file. The warning can be disabled by editing the configuration file and removing it.

Using a specific resolver

Sometimes you may want to use a specific resolver for your project instead of stack init picking one for you. You can do that by using stack init --resolver <resolver>.

You can also init with a compiler resolver if you do not want to use a snapshot. That will result in all of your project's dependencies being put under the extra-deps section.

Installing the compiler

Stack will automatically install the compiler when you run stack build but you can manually specify the compiler by running stack setup <GHC-VERSION>.

Miscellaneous and diagnostics

Add selected packages: If you want to use only selected packages from your project directory you can do so by explicitly specifying the package directories on the command line.

Duplicate package names: If multiple packages under the directory tree have same name, stack init will report those and automatically ignore one of them.

Ignore subdirectories: By default stack init searches all the subdirectories for Cabal files. If you do not want that then you can use --ignore-subdirs command line switch.

Cabal warnings: stack init will show warnings if there were issues in reading a Cabal file. You may want to pay attention to the warnings as sometimes they may result in incomprehensible errors later on during dependency solving.

Package naming: If the Name field defined in a Cabal file does not match with the Cabal file name then stack init will refuse to continue.

User warnings: When packages are excluded or external dependencies added Stack will show warnings every time configuration file is loaded. You can suppress the warnings by editing the config file and removing the warnings from it. You may see something like this:

$ stack build
Warning: Some packages were found to be incompatible with the resolver and have been left commented out in the packages section.
Warning: Specified resolver could not satisfy all dependencies. Some external packages have been added as dependencies.
You can suppress this message by removing it from stack.yaml

Different databases

Time to take a short break from hands-on examples and discuss a little architecture. Stack has the concept of multiple databases.

A database consists of a GHC package database (which contains the compiled version of a library), executables, and a few other things as well. To give you an idea, the contents of the parent directory of the stack path --local-pkg-db directory are the directories:

bin
doc
lib
pkgdb

Databases in Stack are layered. For example, the database listing we just gave is called a local database (also known as a mutable database). That is layered on top of a snapshot database (also known as a write-only database). The snapshot database contains the libraries and executables that are considered to be immutable. Finally, GHC itself ships with a number of libraries and executables, also considered to be immutable, which forms the global database.

To get a quick idea of this, we can look at the output of the stack exec -- ghc-pkg list command in our helloworld project:

<stack path --global-pkg-db directory>
    Cabal-3.6.3.0
    Win32-2.12.0.1
    array-0.5.4.0
    base-4.16.2.0
    binary-0.8.9.0
    bytestring-0.11.3.1
    containers-0.6.5.1
    deepseq-1.4.6.1
    directory-1.3.6.2
    exceptions-0.10.4
    filepath-1.4.2.2
    (ghc-9.2.3)
    ghc-bignum-1.2
    ghc-boot-9.2.3
    ghc-boot-th-9.2.3
    ghc-compact-0.1.0.0
    ghc-heap-9.2.3
    ghc-prim-0.8.0
    ghci-9.2.3
    haskeline-0.8.2
    hpc-0.6.1.0
    integer-gmp-1.1
    libiserv-9.2.3
    mtl-2.2.2
    parsec-3.1.15.0
    pretty-1.1.3.6
    process-1.6.13.2
    rts-1.0.2
    stm-2.5.0.2
    template-haskell-2.18.0.0
    text-1.2.5.0
    time-1.11.1.1
    transformers-0.5.6.2
    xhtml-3000.2.2.1

<stack path --snapshot-pkg-db directory>
    acme-missiles-0.3

<stack path --local-pkg-db directory>
    helloworld-0.1.0.0

where <stack path --global-pkg-db directory> refers to the directory output by the command stack path --global-pkg-db, and so on.

Notice that acme-missiles ends up in the snapshot database. Any package which comes from Hackage, an archive, or a repository is considered to be an immutable package.

Anything which is considered mutable, or depends on something mutable, ends up in the local database. This includes your own code and any other packages located on a local file path.

The reason we have this structure is that:

  • it lets multiple projects reuse the same binary builds of immutable packages,
  • but doesn't allow different projects to "contaminate" each other by putting non-standard content into the shared snapshot database.

As you probably guessed, there can be multiple snapshot databases available. See the contents of the snapshots directory in the Stack root.

  • On Unix-like operating systems, each snapshot is in the last of a sequence of three subdirectories named after the platform, a 256-bit hash of the source map (how the package should be built -- including the compiler, options, and immutable dependencies), and the GHC version.

  • On Windows, each snapshot is in a subdirectory that is a shorter hash (eight characters) of the sequence of three directories used on Unix-like operating systems. This is done to avoid problems created by default limits on file path lengths on Windows systems.

These snapshot databases don't get layered on top of each other; they are each used separately.

In reality, you'll rarely — if ever — interact directly with these databases, but it's good to have a basic understanding of how they work so you can understand why rebuilding may occur at different points.

The build synonyms

Let's look at a subset of the stack --help output:

build    Build the package(s) in this directory/configuration
install  Shortcut for 'build --copy-bins'
test     Shortcut for 'build --test'
bench    Shortcut for 'build --bench'
haddock  Shortcut for 'build --haddock'

Four of these commands are just synonyms for the build command. They are provided for convenience for common cases (e.g., stack test instead of stack build --test) and so that commonly expected commands just work.

What's so special about these commands being synonyms? It allows us to make much more composable command lines. For example, we can have a command that builds executables, generates Haddock documentation (Haskell API-level docs), and builds and runs your test suites, with:

$ stack build --haddock --test

You can even get more inventive as you learn about other flags. For example, take the following:

$ stack build --pedantic --haddock --test --exec "echo Yay, it succeeded" --file-watch

This will:

  • turn on all warnings and errors
  • build your library and executables
  • generate Haddocks
  • build and run your test suite
  • run the command echo Yay, it succeeded when that completes
  • after building, watch for changes in the files used to build the project, and kick off a new build when done

The stack install command and copy-bins option

It's worth calling out the behavior of the install command and --copy-bins option, since this has confused a number of users (especially when compared to behavior of other tools like Cabal (the tool)). The install command does precisely one thing in addition to the build command: it copies any generated executables to the local binary directory. You may recognize the default value for that path:

On Unix-like operating systems:

$ stack path --local-bin
/home/<user_name>/.local/bin

On Windows:

$ stack path --local-bin
C:\Users\<user_name>\AppData\Roaming\local\bin

That's why the download page recommends adding that directory to your PATH. This feature is convenient, because now you can simply run executable-name in your shell instead of having to run stack exec executable-name from inside your project directory.

Since it's such a point of confusion, let me list a number of things Stack does not do specially for the install command:

  • Stack will always build any necessary dependencies for your code. The install command is not necessary to trigger this behavior. If you just want to build a project, run stack build.
  • Stack will not track which files it's copied to your local binary directory nor provide a way to automatically delete them. There are many great tools out there for managing installation of binaries, and Stack does not attempt to replace those.
  • Stack will not necessarily be creating a relocatable executable. If your executables hard-codes paths, copying the executable will not change those hard-coded paths.
    • At the time of writing, there's no way to change those kinds of paths with Stack, but see issue #848 about --prefix for future plans.

That's really all there is to the install command: for the simplicity of what it does, it occupies a much larger mental space than is warranted.

Targets, locals, and extra-deps

We haven't discussed this too much yet, but, in addition to having a number of synonyms and taking a number of options on the command line, the build command also takes many arguments. These are parsed in different ways, and can be used to achieve a high level of flexibility in telling Stack exactly what you want to build.

We're not going to cover the full generality of these arguments here; instead, there's documentation covering the full build command syntax. Here, we'll just point out a few different types of arguments:

  • You can specify a package name, e.g. stack build vector.
    • This will attempt to build the vector package, whether it's a local package, in your extra-deps, in your snapshot, or just available upstream. If it's just available upstream but not included in your locals, extra-deps, or snapshot, the newest version is automatically promoted to an extra-dep.
  • You can also give a package identifier, which is a package name plus version, e.g. stack build yesod-bin-1.4.14.
    • This is almost identical to specifying a package name, except it will (1) choose the given version instead of latest, and (2) error out if the given version conflicts with the version of a local package.
  • The most flexibility comes from specifying individual components, e.g. stack build helloworld:test:helloworld-test says "build the test suite component named helloworld-test from the helloworld package."
    • In addition to this long form, you can also shorten it by skipping what type of component it is, e.g. stack build helloworld:helloworld-test, or even skip the package name entirely, e.g. stack build :helloworld-test.
  • Finally, you can specify individual directories to build to trigger building of any local packages included in those directories or subdirectories.

When you give no specific arguments on the command line (e.g., stack build), it's the same as specifying the names of all of your local packages. If you just want to build the package for the directory you're currently in, you can use stack build ..

Components, --test, and --bench

Here's one final important yet subtle point. Consider our helloworld package: it has a library component, an executable helloworld-exe, and a test suite helloworld-test. When you run stack build helloworld, how does it know which ones to build? By default, it will build the library (if any) and all of the executables but ignore the test suites and benchmarks.

This is where the --test and --bench flags come into play. If you use them, those components will also be included. So stack build --test helloworld will end up including the helloworld-test component as well.

You can bypass this implicit adding of components by being much more explicit, and stating the components directly. For example, the following will not build the helloworld-exe executable once all executables have been sucessfully built:

$ stack clean
$ stack build :helloworld-test
Building all executables for `helloworld' once. After a successful build of all of them, only specified executables will be rebuilt.
helloworld> configure (lib + exe + test)
Configuring helloworld-0.1.0.0...
helloworld> build (lib + exe + test)
Preprocessing library for helloworld-0.1.0.0..
Building library for helloworld-0.1.0.0..
[1 of 2] Compiling Lib
[2 of 2] Compiling Paths_helloworld
Preprocessing executable 'helloworld-exe' for helloworld-0.1.0.0..
Building executable 'helloworld-exe' for helloworld-0.1.0.0..
[1 of 2] Compiling Main
[2 of 2] Compiling Paths_helloworld
Linking .stack-work\dist\<hash>\build\helloworld-exe\helloworld-exe.exe ...
Preprocessing test suite 'helloworld-test' for helloworld-0.1.0.0..
Building test suite 'helloworld-test' for helloworld-0.1.0.0..
[1 of 2] Compiling Main
[2 of 2] Compiling Paths_helloworld
Linking .stack-work\dist\<hash>\build\helloworld-test\helloworld-test.exe ...
helloworld> copy/register
Installing library in ...\helloworld\.stack-work\install\...
Installing executable helloworld-exe in ...\helloworld\.stack-work\install\...\bin
Registering library for helloworld-0.1.0.0..
helloworld> test (suite: helloworld-test)

Test suite not yet implemented

helloworld> Test suite helloworld-test passed
Completed 2 action(s).```

We first cleaned our project to clear old results so we know exactly what Stack is trying to do. Note that is says it is building all executables for helloworld once, and that after a successful build of all of them, only specified executables will be rebuilt. If we change the source code of test/Spec.hs, say to:

main :: IO ()
main = putStrLn "Test suite still not yet implemented"

and command again:

$ stack build :helloworld-test
helloworld-0.1.0.0: unregistering (local file changes: test\Spec.hs)
helloworld> build (lib + test)
Preprocessing library for helloworld-0.1.0.0..
Building library for helloworld-0.1.0.0..
Preprocessing test suite 'helloworld-test' for helloworld-0.1.0.0..
Building test suite 'helloworld-test' for helloworld-0.1.0.0..
[2 of 2] Compiling Main
Linking .stack-work\dist\<hash>\build\helloworld-test\helloworld-test.exe ...
helloworld> copy/register
Installing library in ...\helloworld\.stack-work\install\...
Installing executable helloworld-exe in ...\helloworld\.stack-work\install\...\bin
Registering library for helloworld-0.1.0.0..
helloworld> blocking for directory lock on ...\helloworld\.stack-work\dist\<hash>\build-lock
helloworld> test (suite: helloworld-test)

Test suite still not yet implemented

helloworld> Test suite helloworld-test passed
Completed 2 action(s).

Notice that this time it builds the helloworld-test test suite, and the helloworld library (since it's used by the test suite), but it does not build the helloworld-exe executable.

And now the final point: in both cases, the last line shows that our command also runs the test suite it just built. This may surprise some people who would expect tests to only be run when using stack test, but this design decision is what allows the stack build command to be as composable as it is (as described previously). The same rule applies to benchmarks. To spell it out completely:

  • The --test and --bench flags simply state which components of a package should be built, if no explicit set of components is given
  • The default behavior for any test suite or benchmark component which has been built is to also run it

You can use the --no-run-tests and --no-run-benchmarks flags to disable running of these components. You can also use --no-rerun-tests to prevent running a test suite which has already passed and has not changed.

NOTE: Stack doesn't build or run test suites and benchmarks for non-local packages. This is done so that running a command like stack test doesn't need to run 200 test suites!

Multi-package projects

Until now, everything we've done with Stack has used a single-package project. However, Stack's power truly shines when you're working on multi-package projects. All the functionality you'd expect to work just does: dependencies between packages are detected and respected, dependencies of all packages are just as one cohesive whole, and if anything fails to build, the build commands exits appropriately.

Let's demonstrate this with the wai-app-static and yackage packages, starting in the root directory for all our Haskell projects:

$ mkdir multi
$ cd multi
$ stack unpack wai-app-static yackage
Unpacked wai-app-static (from Hackage) to .../multi/wai-app-static-3.1.7.4/
Unpacked yackage (from Hackage) to .../multi/yackage-0.8.1/
$ stack init
Looking for .cabal or package.yaml files to use to init the project.
Using cabal packages:
- wai-app-static-3.1.7.4/
- yackage-0.8.1/

Cabal file warning in .../multi/yackage-0.8.1/yackage.cabal@47:40: version operators used. To use version operators the package needs to specify at least 'cabal-version: >= 1.8'.
Cabal file warning in .../multi/yackage-0.8.1/yackage.cabal@21:36: version operators used. To use version operators the package needs to specify at least 'cabal-version: >= 1.8'.
Selecting the best among 18 snapshots...

* Matches ...

Selected resolver: ...
Initialising configuration using resolver: ...
Total number of user packages considered: 2
Writing configuration to file: stack.yaml
$ stack build --haddock --test
# Goes off to build a whole bunch of packages

If you look at the stack.yaml file, you'll see exactly what you'd expect:

resolver:
  url: https://raw.githubusercontent.com/commercialhaskell/stackage-snapshots/master/lts/19/17.yaml
packages:
- wai-app-static-3.1.7.4
- yackage-0.8.1

Notice that multiple directories are listed in the packages key.

In addition to local directories, you can also refer to packages available in a Git repository or in a tarball over HTTP/HTTPS. This can be useful for using a modified version of a dependency that hasn't yet been released upstream.

Please note that when adding upstream packages directly to your project it is important to distinguish local packages from the upstream dependency packages. Otherwise you may have trouble running stack ghci. See stack.yaml documentation for more details.

Flags and GHC options

There are two common ways to alter how a package will install: with Cabal flags and with GHC options.

Cabal flag management

To change a Cabal flag setting, we can use the command line --flag option. The yackage package has an upload flag that is enabled by default. We can command:

$ stack build --flag yackage:-upload

This means: when compiling the yackage package, turn off the upload flag (thus the - in -upload). Unlike other tools, Stack is explicit about which package's flag you want to change. It does this for two reasons:

  1. There's no global meaning for Cabal flags, and therefore two packages can use the same flag name for completely different things.
  2. By following this approach, we can avoid unnecessarily recompiling snapshot packages that happen to use a flag that we're using.

You can also change flag values on the command line for extra-dep and snapshot packages. If you do this, that package will automatically be promoted to an extra-dep, since the build plan is different than what the plan snapshot definition would entail.

GHC options

GHC options follow a similar logic as in managing Cabal flags, with a few nuances to adjust for common use cases. Let's consider:

$ stack build --ghc-options="-Wall -Werror"

This will set the -Wall -Werror options for all local targets. Note that this will not affect extra-dep and snapshot packages at all. This design provides us with reproducible and fast builds.

(By the way: the above GHC options have a special convenience flag: --pedantic.)

There's one extra nuance about command line GHC options: Since they only apply to local targets, if you change your local targets, they will no longer apply to other packages. Let's play around with an example from the wai repository, which includes the wai and warp packages, the latter depending on the former. If we run:

$ stack build --ghc-options=-O0 wai

It will build all of the dependencies of wai, and then build wai with all optimizations disabled. Now let's add in warp as well:

$ stack build --ghc-options=-O0 wai warp

This builds the additional dependencies for warp, and then builds warp with optimizations disabled. Importantly: it does not rebuild wai, since wai's configuration has not been altered. Now the surprising case:

$ stack build --ghc-options=-O0 warp
wai-3.0.3.0-5a49351d03cba6cbaf906972d788e65d: unregistering (flags changed from ["--ghc-options","-O0"] to [])
warp-3.1.3-a91c7c3108f63376877cb3cd5dbe8a7a: unregistering (missing dependencies: wai)
wai-3.0.3.0: configure

You may expect this to be a no-op: neither wai nor warp has changed. However, Stack will instead recompile wai with optimizations enabled again, and then rebuild warp (with optimizations disabled) against this newly built wai. The reason: reproducible builds. If we'd never built wai or warp before, trying to build warp would necessitate building all of its dependencies, and it would do so with default GHC options (optimizations enabled). This dependency would include wai. So when we run:

$ stack build --ghc-options=-O0 warp

We want its behavior to be unaffected by any previous build steps we took. While this specific corner case does catch people by surprise, the overall goal of reproducible builds is - in the Stack maintainers' views - worth the confusion.

Final point: if you have GHC options that you'll be regularly passing to your packages, you can add them to your stack.yaml file. See the documentation section on ghc-options for more information.

The stack path command

NOTE: That's it, the heavy content of this guide is done! Everything from here on out is simple explanations of commands. Congratulations!

Generally, you don't need to worry about where Stack stores various files. But some people like to know this stuff. That's when the stack path command is useful. For example:

$ stack path
snapshot-doc-root: ...
local-doc-root: ...
local-hoogle-root: ...
stack-root: ...
project-root: ...
config-location: ...
bin-path: ...
programs: ...
compiler-exe: ...
compiler-bin: ...
compiler-tools-bin: ...
local-bin: ...
extra-include-dirs: ...
extra-library-dirs: ...
snapshot-pkg-db: ...
local-pkg-db: ...
global-pkg-db: ...
ghc-package-path: ...
snapshot-install-root: ...
local-install-root: ...
dist-dir: ...
local-hpc-root: ...
local-bin-path: ...
ghc-paths: ...

In addition, stack path accepts command line arguments to state which of these keys you're interested in, which can be convenient for scripting. As a simple example, let's find out the sandboxed versions of GHC that Stack installed:

On Unix-like operating systems:

$ ls $(stack path --programs)/*.installed
/home/<user_name>/.stack/programs/x86_64-linux/ghc-9.0.2.installed

On Windows (with PowerShell):

$ dir "$(stack path --programs)/*.installed"

    Directory: C:\Users\mikep\AppData\Local\Programs\stack\x86_64-windows

Mode                 LastWriteTime         Length Name
----                 -------------         ------ ----
-a---          27/07/2022  5:40 PM              9 ghc-9.0.2.installed
-a---          25/02/2022 11:39 PM              9 msys2-20210604.installed

While we're talking about paths, to wipe our Stack install completely, here's what needs to be removed:

  1. Delete the Stack root folder (see stack path --stack-root, before you uninstall).
  2. On Windows, delete the folder containing Stack's tools (see stack path --programs, before you uninstall), which is located outside of the Stack root folder
  3. Delete the stack executable (see which stack, on Unix-like operating systems, or where.exe stack, on Windows).

You may also want to delete .stack-work folders in any Haskell projects that you have built using Stack.

The stack exec command

We've already used stack exec multiple times in this guide. As you've likely already guessed, it allows you to run executables, but with a slightly modified environment. In particular: stack exec looks for executables on Stack's bin paths, and sets a few additional environment variables (like adding those paths to the PATH, and setting GHC_PACKAGE_PATH, which tells GHC which package databases to use).

If you want to see exactly what the modified environment looks like, try:

$ stack exec env

The only issue is how to distinguish flags to be passed to Stack versus those for the underlying program. Thanks to the optparse-applicative library, Stack follows the Unix convention of -- to separate these, e.g.:

stack exec --package stm -- echo I installed the stm package via --package stm
Run from outside a project, using implicit global project config
Using latest snapshot resolver: lts-18.3
Writing global (non-project-specific) config file to: /home/michael/.stack/global/stack.yaml
Note: You can change the snapshot via the resolver field there.
I installed the stm package via --package stm

Flags worth mentioning:

  • --package foo can be used to force a package to be installed before running the given command.
  • --no-ghc-package-path can be used to stop the GHC_PACKAGE_PATH environment variable from being set. Some tools — notably Cabal (the tool) — do not behave well with that variable set.

You may also find it convenient to use stack exec to launch a subshell (substitute bash with your preferred shell) where your compiled executable is available at the front of your PATH:

$ stack exec bash

The stack ghci or stack repl command

GHCi is the interactive GHC environment, a.k.a. the REPL. You could access it with:

stack exec ghci

But that won't load up locally written modules for access. For that, use the stack ghci command or its synonym stack repl. To then load modules from your project, use the :m command (for "module") followed by the module name.

IMPORTANT NOTE: If you have added upstream packages to your project please make sure to mark them as dependency packages for faster and reliable usage of stack ghci. Otherwise GHCi may have trouble due to conflicts of compilation flags or having to unnecessarily interpret too many modules. See stack.yaml documentation to learn how to mark a package as a dependency package.

The stack ghc and stack runghc commands

You'll sometimes want to just compile (or run) a single Haskell source file, instead of creating an entire Cabal package for it. You can use stack exec ghc or stack exec runghc for that. As simple helpers, we also provide the stack ghc and stack runghc commands, for these common cases.

The script interpreter and stack script command

Stack also offers a very useful feature for running files: a script interpreter. For too long have Haskellers felt shackled to bash or Python because it's just too hard to create reusable source-only Haskell scripts. Stack attempts to solve that.

You can use stack <file name> to execute a Haskell source file or specify Stack as the interpreter using a shebang line on a Unix like operating systems. Additional Stack options can be specified using a special Haskell comment in the source file to specify dependencies and automatically install them before running the file.

An example will be easiest to understand. Consider the Haskell source file turtle-example.hs with contents:

#!/usr/bin/env stack
-- stack script --resolver lts-19.17 --package turtle
{-# LANGUAGE OverloadedStrings #-}
import Turtle
main = echo "Hello World!"

On Unix-like operating systems the file's permissions can changed to mark it as executable, with command chmod and then it can be run:

$ chmod +x turtle-example.hs
$ ./turtle-example.hs

The first line in the source file beginning with the 'shebang' (#!) tells Unix to use Stack as a script interpreter.

On Windows, PowerShell will not recognise the shebang line, and so the line is not required on Windows. However, the script can be run with command:

$ stack turtle-example.hs

In both cases, the command yields (first time):

# Progress with building dependencies...
Hello World!

and (second and subsequent times):

Hello World!

The first run can take a while (as it has to download GHC if necessary and build dependencies), but subsequent runs are able to reuse everything already built, and are therefore quite fast.

The second line of the source code is a Haskell comment providing additional options to Stack. (A 'shebang' line is limited to a single argument). In this example, the options tell Stack to use the LTS Haskell 19.17 resolver and ensure the turtle package is available.

Just-in-time compilation

You can add the --compile flag to make Stack compile the script, and then run the compiled executable. Compilation is done quickly, without optimization. To compile with optimization, use the --optimize flag instead. Compilation is done only if needed; if the executable already exists, and is newer than the script, Stack just runs the executable directly.

This feature can be good for speed (your script runs faster) and also for durability (the executable remains runnable even if the script is disturbed, eg due to changes in your installed GHC/snapshots, changes to source files during git bisect, etc.)

Using multiple packages

You can also specify multiple packages, either with multiple --package arguments, or by providing a comma or space separated list. For example:

#!/usr/bin/env stack
{- stack
  script
  --resolver lts-19.17
  --package turtle
  --package "stm async"
  --package http-client,http-conduit
-}

Stack configuration for scripts

With the script command, all Stack configuration files are ignored to provide a completely reliable script running experience. However, see the example below with runghc for an approach to scripts which will respect your configuration files. When using runghc, if the current working directory is inside a project then that project's Stack configuration is effective when running the script. Otherwise the script uses the global project configuration specified in <Stack root>/global-project/stack.yaml.

Specifying interpreter options

The Stack interpreter options comment must specify a single valid Stack command line, starting with stack as the command followed by the Stack options to use for executing this file. The comment must always be on the line immediately following the shebang line when the shebang line is present otherwise it must be the first line in the file. The comment must always start in the first column of the line.

When many options are needed a block style comment may be more convenient to split the command on multiple lines for better readability. You can also specify GHC options the same way as you would on command line i.e. by separating the Stack options and GHC options with a --. Here is an example of a multi line block comment with GHC options:

#!/usr/bin/env stack
{- stack
  script
  --resolver lts-19.17
  --package turtle
  --
  +RTS -s -RTS
-}

Writing independent and reliable scripts

The stack script command will automatically:

  • Install GHC and libraries if missing
  • Require that all packages used be explicitly stated on the command line

This ensures that your scripts are independent of any prior deployment specific configuration, and are reliable by using exactly the same version of all packages every time it runs so that the script does not break by accidentally using incompatible package versions.

In earlier versions of Stack, the runghc command was used for scripts and can still be used in that way. In order to achieve the same effect with the runghc command, you can do the following:

  1. Use the --install-ghc option to install the compiler automatically
  2. Explicitly specify all packages required by the script using the --package option. Use -hide-all-packages GHC option to force explicit specification of all packages.
  3. Use the --resolver Stack option to ensure a specific GHC version and package set is used.

It is possible for configuration files to affect stack runghc. For that reason, stack script is strongly recommended. For those curious, here is an example with runghc:

#!/usr/bin/env stack
{- stack
  runghc
  --install-ghc
  --resolver lts-19.17
  --package base
  --package turtle
  --
  -hide-all-packages
  -}

The runghc command is still very useful, especially when you're working on a project and want to access the package databases and configurations used by that project. See the next section for more information on configuration files.

Platform-specific script issues

On macOS:

  • Avoid {-# LANGUAGE CPP #-} in Stack scripts; it breaks the shebang line (GHC #6132)

  • Use a compiled executable, not another script, in the shebang line. Eg #!/usr/bin/env runhaskell will work but #!/usr/local/bin/runhaskell would not.

Loading scripts in GHCi

Sometimes you want to load your script in GHCi to play around with your program. In those cases, you can use exec ghci option in the script to achieve it. Here is an example:

#!/usr/bin/env stack
{- stack
   exec ghci
   --install-ghc
   --resolver lts-19.17
   --package turtle
-}

Finding project configs, and the implicit global project

Whenever you run something with Stack, it needs a project-level configuration file. The algorithm Stack uses to find such a file is:

  1. Check for a --stack-yaml option on the command line
  2. Check for a STACK_YAML environment variable
  3. Check the current directory and all ancestor directories for a stack.yaml file

The first two provide a convenient method for using an alternate configuration. For example: stack build --stack-yaml stack-ghc-9.2.3.yaml can be used by your CI system to check your code against GHC 9.2.3. Setting the STACK_YAML environment variable can be convenient if you're going to be running commands like stack ghc in other directories, but you want to use the configuration you defined in a specific project.

If Stack does not find a project level configuration file in any of the three specified locations, the implicit global logic kicks in. You've probably noticed that phrase a few times in the output from commands above. Implicit global is essentially a hack to allow Stack to be useful in a non-project setting. When no implicit global configuration file exists, Stack creates one for you with the latest LTS snapshot as the resolver. This allows you to do things like:

  • compile individual files easily with stack ghc
  • build executables without starting a project, e.g. stack install pandoc

Keep in mind that there's nothing magical about this implicit global configuration. It has no effect on projects at all. Every package you install with it is put into isolated databases just like everywhere else. The only magic is that it's the catch-all project whenever you're running Stack somewhere else.

Setting the Stack root location

The Stack root is a directory where Stack stores snapshot packages. On Unix-like operating systems, it is also where Stack stores tools such as GHC, in a programs directory. On Windows, the default location for such tools is outside the Stack root, in %LOCALAPPDATA%\Programs\stack. The location of the Stack root is reported by command:

$ stack path --stack-root

The location of the Stack root can be configured by setting the STACK_ROOT environment variable or using Stack's --stack-root option on the command line.

On Windows, the length of filepaths may be limited (to MAX_PATH), and things can break when this limit is exceeded. Setting a Stack root with a short path to its location (for example, C:\sr) can help.

stack.yaml versus Cabal files

Now that we've covered a lot of Stack use cases, this quick summary of stack.yaml versus Cabal files will hopefully make sense and be a good reminder for future uses of Stack:

  • A project can have multiple packages.
  • Each project has a stack.yaml.
  • Each package has a Cabal file, named <package_name>.cabal.
  • The Cabal file specifies which packages are dependencies.
  • The stack.yaml file specifies which packages are available to be used.
  • The Cabal file specifies the components, modules, and build flags provided by a package
  • stack.yaml can override the flag settings for individual packages
  • stack.yaml specifies which packages to include

Comparison to other tools

Stack is not the only tool around for building Haskell code. Stack came into existence due to limitations with some of the existing tools. If you're unaffected by those limitations and are happily building Haskell code, you may not need Stack. If you're suffering from some of the common problems in other tools, give Stack a try instead.

If you're a new user who has no experience with other tools, we recommend going with Stack. The defaults match modern best practices in Haskell development, and there are less corner cases you need to be aware of. You can develop Haskell code with other tools, but you probably want to spend your time writing code, not convincing a tool to do what you want.

Before jumping into the differences, let me clarify an important similarity:

Same package format. Stack, Cabal (the tool), and presumably all other tools share the same underlying Cabal package format, consisting of a Cabal file, modules, etc. This is a Good Thing: we can share the same set of upstream libraries, and collaboratively work on the same project with Stack, Cabal (the tool), and NixOS. In that sense, we're sharing the same ecosystem.

Now the differences:

  • Curation vs dependency solving as a default.
    • Stack defaults to using curation (Stackage snapshots, LTS Haskell, Nightly, etc) as a default instead of defaulting to dependency solving, as Cabal (the tool) does. This is just a default: as described above, Stack can use dependency solving if desired, and Cabal (the tool) can use curation. However, most users will stick to the defaults. The Stack team firmly believes that the majority of users want to simply ignore dependency resolution nightmares and get a valid build plan from day one, which is why we've made this selection of default behavior.
  • Reproducible.
    • Stack goes to great lengths to ensure that stack build today does the same thing tomorrow. Cabal (the tool) does not: build plans can be affected by the presence of pre-installed packages, and running cabal update can cause a previously successful build to fail. With Stack, changing the build plan is always an explicit decision.
  • Automatically building dependencies.
    • With Cabal (the tool), you need to use cabal install to trigger dependency building. This is somewhat necessary due to the previous point, since building dependencies can, in some cases, break existing installed packages. So for example, in Stack, stack test does the same job as cabal install --run-tests, though the latter additionally performs an installation that you may not want. The closer equivalent command sequence is: cabal install --enable-tests --only-dependencies, cabal configure --enable-tests, cabal build && cabal test (newer versions of Cabal (the tool) may make this command sequence shorter).
  • Isolated by default.
    • This has been a pain point for new Stack users. In Cabal, the default behavior is a non-isolated build where working on two projects can cause the user package database to become corrupted. The Cabal solution to this is sandboxes. Stack, however, provides this behavior by default via its databases. In other words: when you use Stack, there's no need for sandboxes, everything is (essentially) sandboxed by default.

Other tools for comparison (including active and historical)

  • cabal-dev. This is deprecated in favor of Cabal (the tool).
  • cabal-meta inspired a lot of the multi-package functionality of Stack. If you're still using Cabal (the tool), cabal-meta is relevant. For Stack work, the feature set is fully subsumed by Stack.
  • cabal-src is mostly irrelevant in the presence of both Stack and Cabal sandboxes, both of which make it easier to add additional package sources easily. The mega-sdist executable that ships with cabal-src is, however, still relevant. Its functionality may some day be folded into Stack
  • stackage-cli was an initial attempt to make Cabal (the tool) work more easily with curated snapshots, but due to a slight impedance mismatch between cabal.config constraints and snapshots, it did not work as well as hoped. It is deprecated in favor of Stack.

Fun features

This is just a quick collection of fun and useful feature Stack supports.

Templates

We started off using the new command to create a project. Stack provides multiple templates to start a new project from. For more information, command:

$ stack templates

You can specify one of these templates following your project name in the stack new command:

$ stack new my-rio-project rio
Downloading template "rio" to create project "my-rio-project" in my-rio-project/ ...
Looking for .cabal or package.yaml files to use to init the project.
Using cabal packages:
- my-rio-project/

Selecting the best among 18 snapshots...

* Matches ...

Selected resolver: ...
Initialising configuration using resolver: ...
Total number of user packages considered: 1
Writing configuration to file: my-rio-project/stack.yaml
All done.
<Stack root>\templates\rio.hsfiles:   10.10 KiB downloaded...

The default templates repository is https://github.com/commercialhaskell/stack-templates. You can download templates from a different Github user by prefixing the username and a slash:

$ stack new my-yesod-project yesodweb/simple

Then template file simple.hsfiles would be downloaded from GitHub repository yesodweb/stack-templates.

You can even download templates from a service other that GitHub, such as GitLab or Bitbucket. For example:

$ stack new my-project gitlab:user29/foo

Template file foo.hsfiles would be downloaded from GitLab, user account user29, repository stack-templates.

If you need more flexibility, you can specify the full URL of the template:

$ stack new my-project https://my-site.com/content/template9.hsfiles

(The .hsfiles extension is optional; it will be added if it's not specified.)

Alternatively you can use a local template by specifying the path:

$ stack new project <path_to_template>/template.hsfiles

As a starting point for creating your own templates, you can use the "simple" template. The stack-templates repository provides an introduction into creating templates.

Editor integration

For editor integration, Stack has a related project called intero. It is particularly well supported by Emacs, but some other editors have integration for it as well.

Visualizing dependencies

If you'd like to get some insight into the dependency tree of your packages, you can use the stack dot command and Graphviz. More information is available in the Dependency visualization documentation.

Travis with caching

This content has been moved to a dedicated Travis CI document.

Shell auto-completion

Love tab-completion of commands? You're not alone. If you're on bash, just run the following (or add it to .bashrc):

eval "$(stack --bash-completion-script stack)"

For more information and other shells, see the Shell auto-completion wiki page

Docker

Stack is able to build your code inside a Docker image, which means even more reproducibility to your builds, since you and the rest of your team will always have the same system libraries.

Nix

Stack provides an integration with Nix, providing you with the same two benefits as the first Docker integration discussed above:

  • more reproducible builds, since fixed versions of any system libraries and commands required to build the project are automatically built using Nix and managed locally per-project. These system packages never conflict with any existing versions of these libraries on your system. That they are managed locally to the project means that you don't need to alter your system in any way to build any odd project pulled from the Internet.
  • implicit sharing of system packages between projects, so you don't have more copies on-disk than you need to.

When using the Nix integration, Stack downloads and builds Haskell dependencies as usual, but resorts on Nix to provide non-Haskell dependencies that exist in the Nixpkgs.

Both Docker and Nix are methods to isolate builds and thereby make them more reproducible. They just differ in the means of achieving this isolation. Nix provides slightly weaker isolation guarantees than Docker, but is more lightweight and more portable (Linux and macOS mainly, but also Windows). For more on Nix, its command-line interface and its package description language, read the Nix manual. But keep in mind that the point of Stack's support is to obviate the need to write any Nix code in the common case or even to learn how to use the Nix tools (they're called under the hood).

For more information, see the Nix-integration documentation.

Power user commands

The following commands are a little more powerful, and won't be needed by all users. Here's a quick rundown:

  • stack update will download the most recent set of packages from your package indices (e.g. Hackage). Generally, Stack runs this for you automatically when necessary, but it can be useful to do this manually sometimes.
  • stack unpack is a command we've already used quite a bit for examples, but most users won't use it regularly. It does what you'd expect: downloads a tarball and unpacks it. It accept optional --to argument to specify the destination directory.
  • stack sdist generates an uploading tarball containing your package code
  • stack upload uploads an sdist to Hackage. As of version 1.1.0 Stack will also attempt to GPG sign your packages as per our blog post.

    • --no-signature disables signing of packages
    • --candidate upload a package candidate
    • Hackage API key can be used instead of username and password. Usage example:

    bash HACKAGE_KEY=<api_key> stack upload .

    • username and password can be read by environment

    bash export $HACKAGE_USERNAME="<username>" export $HACKAGE_PASSWORD="<password>"

  • stack upgrade will build a new version of Stack from source.

    • --git is a convenient way to get the most recent version from the master branch, for those testing and living on the bleeding edge.
  • stack ls snapshots will list all the local snapshots by default. You can also view the remote snapshots using stack ls snapshots remote. It also supports option for viewing only lts (-l) and nightly (-n) snapshots.
  • stack ls dependencies lists all of the packages and versions used for a project
  • stack list [PACKAGE]... list the version of the specified package(s) in a snapshot, or without an argument list all the snapshot's package versions. If no resolver is specified the latest package version from Hackage is given.
  • stack sig subcommand can help you with GPG signing & verification
    • sign will sign an sdist tarball and submit the signature to sig.commercialhaskell.org for storage in the sig-archive git repo. (Signatures will be used later to verify package integrity.)

Debugging

To profile a component of the current project, simply pass the --profile flag to stack. The --profile flag turns on the --enable-library-profiling and --enable-executable-profiling Cabal options and passes the +RTS -p runtime options to any testsuites and benchmarks.

For example the following command will build the my-tests testsuite with profiling options and create a my-tests.prof file in the current directory as a result of the test run.

$ stack test --profile my-tests

The my-tests.prof file now contains time and allocation info for the test run.

To create a profiling report for an executable, e.g. my-exe, you can run

 $ stack exec --profile -- my-exe +RTS -p

For more fine-grained control of compilation options there are the --library-profiling and --executable-profiling flags which will turn on the --enable-library-profiling and --enable-executable-profiling Cabal options respectively. Custom GHC options can be passed in with --ghc-options "more options here".

To enable compilation with profiling options by default you can add the following snippet to your stack.yaml or ~/.stack/config.yaml:

build:
  library-profiling: true
  executable-profiling: true

Further reading

For more commands and uses, see the official GHC chapter on profiling, the Haskell wiki, and the chapter on profiling in Real World Haskell.

Tracing

To generate a backtrace in case of exceptions during a test or benchmarks run, use the --trace flag. Like --profile this compiles with profiling options, but adds the +RTS -xc runtime option.

Debugging symbols

Building with debugging symbols in the DWARF information is supported by Stack. This can be done by passing the flag --ghc-options="-g" and also to override the default behaviour of stripping executables of debugging symbols by passing either one of the following flags: --no-strip, --no-library-stripping or --no-executable-stripping.

In Windows, GDB can be installed to debug an executable with stack exec -- pacman -S gdb. Windows' Visual Studio compiler's debugging format PDB is not supported at the moment. This might be possible by separating debugging symbols and converting their format. Or as an option when using the LLVM backend.

More resources

There are lots of resources available for learning more about Stack: