Embed a single file in your source code.
import qualified Data.ByteString
myFile :: Data.ByteString.ByteString
myFile = $(embedFile "dirName/fileName"):: a typeCtrl KGHC 9.10.3 · lts/ghc-9.10.x · c74966e · 2026-09-27
Modulefile-embed-0.0.16.0Haskell2010
This module uses template Haskell. Following is a simplified explanation of usage for those unfamiliar with calling Template Haskell functions.
The function embedFile in this modules embeds a file into the executable
that you can use it at runtime. A file is represented as a ByteString.
However, as you can see below, the type signature indicates a value of type
Q Exp will be returned. In order to convert this into a ByteString, you
must use Template Haskell syntax, e.g.:
$(embedFile "myfile.txt")This expression will have type ByteString. Be certain to enable the
TemplateHaskell language extension, usually by adding the following to the
top of your module:
{-# LANGUAGE TemplateHaskell #-}Embed a single file in your source code.
import qualified Data.ByteString
myFile :: Data.ByteString.ByteString
myFile = $(embedFile "dirName/fileName")Embed a single file in your source code. Unlike embedFile, path is given relative to project root. @since 0.0.16.0
Maybe embed a single file in your source code depending on whether or not file exists.
Warning: When a build is compiled with the file missing, a recompile when the file exists might not trigger an embed of the file. You might try to fix this by doing a clean build.
import qualified Data.ByteString
maybeMyFile :: Maybe Data.ByteString.ByteString
maybeMyFile = $(embedFileIfExists "dirName/fileName")Embed a single existing file in your source code out of list a list of paths supplied.
import qualified Data.ByteString
myFile :: Data.ByteString.ByteString
myFile = $(embedOneFileOf [ "dirName/fileName", "src/dirName/fileName" ])Embed a directory recursively in your source code.
import qualified Data.ByteString
myDir :: [(FilePath, Data.ByteString.ByteString)]
myDir = $(embedDir "dirName")Embed a directory listing recursively in your source code.
myFiles :: [FilePath]
myFiles = $(embedDirListing "dirName")Get a directory tree in the IO monad.
This is the workhorse of embedDir
Embed a single file in your source code.
import Data.String
myFile :: IsString a => a
myFile = $(embedStringFile "dirName/fileName")Since 0.0.9
Embed a single existing string file in your source code out of list a list of paths supplied.
Since 0.0.9
The inject system allows arbitrary content to be embedded inside a Haskell executable, post compilation. Typically, file-embed allows you to read some contents from the file system at compile time and embed them inside your executable. Consider a case, instead, where you would want to embed these contents after compilation. Two real-world examples are:
You would like to embed a hash of the executable itself, for sanity checking in a network protocol. (Obviously the hash will change after you embed the hash.)
You want to create a self-contained web server that has a set of content, but will need to update the content on machines that do not have access to GHC.
The typical workflow use:
Use dummySpace or dummySpaceWith to create some empty space in your executable
Use injectFile or injectFileWith from a separate utility to modify that executable to have the updated content.
The reason for the With-variant of the functions is for cases where you wish
to inject multiple different kinds of content, and therefore need control over
the magic key. If you know for certain that there will only be one dummy space
available, you can use the non-With variants.
Allocate the given number of bytes in the generate executable. That space can be filled up with the inject and injectFile functions.
Like dummySpace, but takes a postfix for the magic string. In
order for this to work, the same postfix must be used by inject /
injectFile. This allows an executable to have multiple
ByteStrings injected into it, without encountering collisions.
Since 0.0.8
inject :: ByteStringbs to inject
-> ByteStringoriginal BS containing dummy
-> Maybe ByteStringnew BS, or Nothing if there is insufficient dummy space
Inject some raw data inside a ByteString containing empty, dummy space
(allocated with dummySpace). Typically, the original ByteString is an
executable read from the filesystem.
Same as inject, but instead of performing the injecting in memory, read the contents from the filesystem and write back to a different file on the filesystem.
injectWith :: ByteStringpostfix of magic string
-> ByteStringbs to inject
-> ByteStringoriginal BS containing dummy
-> Maybe ByteStringnew BS, or Nothing if there is insufficient dummy space
Like inject, but takes a postfix for the magic string.
Since 0.0.8
injectFileWith :: ByteStringpostfix of magic string
-> ByteStringbs to inject
-> FilePathtemplate file
-> FilePathoutput file
-> IO ()Like injectFile, but takes a postfix for the magic string.
Since 0.0.8
Take a relative file path and attach it to the root of the current project.
The idea here is that, when building with Stack, the build will always be
executed with a current working directory of the root of the project (where
your .cabal file is located). However, if you load up multiple projects with
stack ghci, the working directory may be something else entirely.
This function looks at the source location of the Haskell file calling it, finds the first parent directory with a .cabal file, and uses that as the root directory for fixing the relative path.
$(makeRelativeToProject "data/foo.txt" >>= embedFile)Take a predicate to infer the project root and a relative file path, the given file path is then attached to the inferred project root
This function looks at the source location of the Haskell file calling it, finds the first parent directory with a file matching the given predicate, and uses that as the root directory for fixing the relative path.
$(makeRelativeToLocationPredicate ((==) ".cabal" . takeExtension) "data/foo.txt" >>= embedFile)