shake-0.9: Build system library, like Make, but more accurate dependencies.

Development.Shake

Description

This module is used for defining Shake build systems. As a simple example of a Shake build system, let us build the file result.tar from the files listed by result.txt:

import Development.Shake
import Development.Shake.FilePath

main = shake shakeOptions $do want ["result.tar"] "*.tar" *> \out -> do contents <- readFileLines$ replaceExtension out "txt"
need contents
system' "tar" $["-cf",out] ++ contents  We start by importing the modules defining both Shake and routines for manipulating FilePath values. We define main to call shake with the default shakeOptions. As the second argument to shake, we provide a set of rules. There are two common forms of rules, want to specify target files, and *> to define a rule which builds a FilePattern. We use want to require that after the build completes the file result.tar should be ready. The *.tar rule describes how to build files with the extension .tar, including result.tar. We readFileLines on result.txt, after changing the .tar extension to .txt. We read each line into the variable contents -- being a list of the files that should go into result.tar. Next, we depend (need) all the files in contents. If any of these files change, the rule will be repeated. Finally we call the tar program. If either result.txt changes, or any of the files listed by result.txt change, then result.tar will be rebuilt. When writing a Shake build system, start by defining what you want, then write rules with *> to produce the results. Before calling system' you should ensure that any files the command requires are demanded with calls to need. We offer the following advice to Shake users: • If ghc --make or cabal is capable of building your project, use that instead. Custom build systems are necessary for many complex projects, but many projects are not complex. • The CmdArgs package (http://hackage.haskell.org/package/cmdargs/) is well suited to providing command line parsing for build systems, often using flags to set fields in shakeOptions. • Put all result files in a distinguished directory, for example _make. You can implement a clean command by removing that directory, using removeDirectoryRecursive. • To obtain parallel builds set shakeThreads to a number greater than 1. You may also need to compile with -threaded. • Often the want commands will be determined by command line arguments, to mirror the behaviour of make targets. For a default set of want commands that you later override, withoutActions can be useful. • Lots of compilers produce .o files. To avoid overlapping rules, use .c.o for C compilers, .hs.o for Haskell compilers etc. • Do not be afraid to mix Shake rules, system commands and other Haskell libraries -- use each for what it does best. • The more accurate the dependencies are, the better. Use additional rules like doesFileExist and getDirectoryFiles to track information other than just the contents of files. For information in the environment that you suspect will change regularly (perhaps ghc version number), either write the information to a file with alwaysRerun and writeFileChanged, or use addOracle. The theory behind Shake is described in an ICFP 2012 paper, Shake Before Building -- Replacing Make with Haskell http://community.haskell.org/~ndm/downloads/paper-shake_before_building-10_sep_2012.pdf. The associated talk forms a short overview of Shake http://www.youtube.com/watch?v=xYCPpXVlqFM. Acknowledgements: Thanks to Austin Seipp for properly integrating the profiling code. Synopsis # Core shake :: ShakeOptions -> Rules () -> IO ()Source Main entry point for running Shake build systems. For an example see the top of the module Development.Shake. Use ShakeOptions to specify how the system runs, and Rules to specify what to build. The function will throw an exception if the build fails. To use command line flags to modify ShakeOptions see shakeWithArgs. The default set of ShakeOptions. type ShakeValue a = (Show a, Typeable a, Eq a, Hashable a, Binary a, NFData a)Source Define an alias for the six type classes required for things involved in Shake Rules. This alias is only available in GHC 7.4 and above, and requires the ConstraintKinds extension. To define your own values meeting the necessary constraints it is convenient to use the extensions GeneralizedNewtypeDeriving and DeriveDataTypeable to write:  newtype MyType = MyType (String, Bool) deriving (Show,Typeable,Eq,Hashable,Binary,NFData)  class (ShakeValue key, ShakeValue value) => Rule key value | key -> value whereSource Define a pair of types that can be used by Shake rules. To import all the type classes required see Development.Shake.Classes. Methods storedValue :: key -> IO (Maybe value)Source Retrieve the value associated with a key, if available. As an example for filenames/timestamps, if the file exists you should return Just the timestamp, but otherwise return Nothing. For rules whose values are not stored externally, storedValue should return Nothing. data Rules a Source Define a set of rules. Rules can be created with calls to rule, defaultRule or action. Rules are combined with either the Monoid instance, or (more commonly) the Monad instance and do notation. Instances  Monad Rules Functor Rules Applicative Rules Monoid a => Monoid (Rules a) defaultRule :: Rule key value => (key -> Maybe (Action value)) -> Rules ()Source Like rule, but lower priority, if no rule exists then defaultRule is checked. All default rules must be disjoint. rule :: Rule key value => (key -> Maybe (Action value)) -> Rules ()Source Add a rule to build a key, returning an appropriate Action. All rules must be disjoint. To define lower priority rules use defaultRule. action :: Action a -> Rules ()Source Run an action, usually used for specifying top-level requirements. withoutActions :: Rules () -> Rules ()Source Remove all actions specified in a set of rules, usually used for implementing command line specification of what to build. data Action a Source The Action monad, use liftIO to raise IO actions into it, and need to execute files. Action values are used by rule and action. Instances  Monad Action Functor Action Applicative Action MonadIO Action apply :: Rule key value => [key] -> Action [value]Source Execute a rule, returning the associated values. If possible, the rules will be run in parallel. This function requires that appropriate rules have been added with rule or defaultRule. apply1 :: Rule key value => key -> Action valueSource Apply a single rule, equivalent to calling apply with a singleton list. Where possible, use apply to allow parallelism. traced :: String -> IO a -> Action aSource Write an action to the trace list, along with the start/end time of running the IO action. The system' command automatically calls traced. The trace list is used for profile reports (see shakeReport). liftIO :: MonadIO m => forall a. IO a -> m a Lift a computation from the IO monad. # Configuration Options to control the execution of Shake, usually specified by overriding fields in shakeOptions:  shakeOptions{shakeThreads=4, shakeReport=Just "report.html"} The Data instance for this type reports the shakeProgress field as having the abstract type Function, because Data cannot be defined for functions. Constructors  ShakeOptions FieldsshakeFiles :: FilePathDefaults to .shake. The prefix of the filename used for storing Shake metadata files. All metadata files will be named shakeFiles.extension, for some extension. shakeThreads :: IntDefaults to 1. Maximum number of rules to run in parallel, similar to make --jobs=N. To enable parallelism you may need to compile with -threaded. For many build systems, a number equal to or slightly less than the number of physical processors works well. shakeVersion :: StringDefaults to 1. The version number of your build rules. Change the version number to force a complete rebuild, such as when making significant changes to the rules that require a wipe. The version number should be set in the source code, and not passed on the command line. shakeVerbosity :: VerbosityDefaults to Normal. What level of messages should be printed out. shakeStaunch :: BoolDefaults to False. Operate in staunch mode, where building continues even after errors, similar to make --keep-going. shakeReport :: Maybe FilePathDefaults to Nothing. Write an HTML profiling report to a file, showing which rules rebuilt, why, and how much time they took. Useful for improving the speed of your build systems. shakeLint :: BoolDefaults to False. Perform basic sanity checks during building, checking the current directory is not modified and that output files are not modified by multiple rules. These sanity checks do not check for missing or redundant dependencies. shakeDeterministic :: BoolDefaults to False. Run rules in a deterministic order, as far as possible. Typically used in conjunction with shakeThreads=1 for reproducing a build. If this field is set to False, Shake will run rules in a random order, which typically decreases contention for resources and speeds up the build. shakeFlush :: Maybe DoubleDefaults to Just 10. How often to flush Shake metadata files in seconds, or Nothing to never flush explicitly. It is possible that on abnormal termination (not Haskell exceptions) any rules that completed in the last shakeFlush seconds will be lost. shakeAssume :: Maybe AssumeDefaults to Nothing. Assume all build objects are clean/dirty, see Assume for details. Can be used to implement make --touch. shakeAbbreviations :: [(String, String)]Defaults to []. A list of substrings that should be abbreviated in status messages, and their corresponding abbreviation. Commonly used to replace the long paths (e.g. .make/i586-linux-gcc/output) with an abbreviation (e.g. $OUT). shakeStorageLog :: BoolDefaults to False. Write a message to shakeFiles.storage whenever a storage event happens which may impact on the current stored progress. Examples include database version number changes, database compaction or corrupt files. shakeProgress :: IO Progress -> IO ()Defaults to no action. A function called on a separate thread when the build starts, allowing progress to be reported. For applications that want to display progress messages, progressSimple is often sufficient, but more advanced users should look at the Progress data type. shakeOutput :: Verbosity -> String -> IO ()Defaults to writing using putStrLn. A function called to output messages from Shake, along with the Verbosity at which that message should be printed. This function will be called atomically from all other shakeOutput functions. The Verbosity will always be greater than or higher than shakeVerbosity.

Instances

 Data ShakeOptions Show ShakeOptions Typeable ShakeOptions

data Assume Source

The current assumptions made by the build system, used by shakeAssume. These options allow the end user to specify that any rules run are either to be treated as clean, or as dirty, regardless of what the build system thinks.

These assumptions only operate on files reached by the current action commands. Any other files in the database are left unchanged.

Constructors

 AssumeDirty Assume that all rules reached are dirty and require rebuilding, equivalent to storedValue always returning Nothing. Useful to undo the results of AssumeClean, for benchmarking rebuild speed and for rebuilding if untracked dependencies have changed. This assumption is safe, but may cause more rebuilding than necessary. AssumeClean This assumption is unsafe, and may lead to incorrect build results. Assume that all rules reached are clean and do not require rebuilding, provided the rule has a storedValue and has been built before. Useful if you have modified a file in some inconsequential way, such as only the comments or whitespace, and wish to avoid a rebuild.

Instances

 Bounded Assume Enum Assume Eq Assume Data Assume Ord Assume Show Assume Typeable Assume

## Command line

shakeWithArgs :: IO () -> ShakeOptions -> Rules () -> IO ()Source

Run a build system using command line arguments for configuration. Requires a way of cleaning the build objects (triggered by clean as a target), a base set of options that may be overriden by command line flags, and the set of build rules. The function removeFiles is often useful for producing a cleaning action.

The available command line options are those from shakeOptDescrs, along with a few additional make compatible flags that are not represented in ShakeOptions, such as --print-directory.

A list of command line options that can be used to modify ShakeOptions. Each option returns either an error message (invalid argument to the flag) or a function that changes some fields in ShakeOptions. The command line flags are make compatible where possbile, but additional flags have been added for the extra options Shake supports.

removeFiles :: FilePath -> [FilePattern] -> IO ()Source

Remove all empty directories and files that match any of the patterns beneath a directory. Some examples:

   removeFiles "output" ["//*"]
removeFiles "." ["//*.hi","//*.o"]


This function is often useful when writing a 'clean' function for your build system, often as the first argument to shakeWithArgs.

## Progress reporting

data Progress Source

Information about the current state of the build, obtained by passing a callback function to shakeProgress. Typically a program will use progressDisplay to poll this value and produce status messages, which is implemented using this data type.

Constructors

 Progress FieldsisRunning :: !BoolStarts out True, becomes False once the build has completed. countSkipped :: !IntNumber of rules which were required, but were already in a valid state. countBuilt :: !IntNumber of rules which were have been built in this run. countUnknown :: !IntNumber of rules which have been built previously, but are not yet known to be required. countTodo :: !IntNumber of rules which are currently required (ignoring dependencies that do not change), but not built. timeSkipped :: !DoubleTime spent building countSkipped rules in previous runs. timeBuilt :: !DoubleTime spent building countBuilt rules. timeUnknown :: !DoubleTime spent building countUnknown rules in previous runs. timeTodo :: !(Double, Int)Time spent building countTodo rules in previous runs, plus the number which have no known time (have never been built before).

Instances

 Eq Progress Data Progress Ord Progress Show Progress Typeable Progress Monoid Progress

A simple method for displaying progress messages, suitable for using as shakeProgress. This function writes the current progress to the titlebar every five seconds. The function is defined as:

progressSimple = progressDisplay 5 progressTitlebar


progressDisplay :: Double -> (String -> IO ()) -> IO Progress -> IO ()Source

Given a sampling interval (in seconds) and a way to display the status message, produce a function suitable for using as shakeProgress. This function polls the progress information every n seconds, produces a status message and displays it using the display function.

Typical status messages will take the form of 1:25m (15%), indicating that the build is predicted to complete in 1min 25sec, and 15% of the necessary build time has elapsed. This function uses past observations to predict future behaviour, and as such, is only guessing. The time is likely to go up as well as down, and will be less accurate from a clean build (as the system has fewer past observations).

The current implementation is to predict the time remaining (based on timeTodo) and the work already done (timeBuilt). The percentage is then calculated as remaining / (done + remaining), while time left is calculated by scaling remaining by the observed work rate in this build, namely done / time_elapsed.

Set the title of the current console window to the given text. On Windows this function uses the SetConsoleTitle API, elsewhere it uses an xterm escape sequence. This function may not work for all terminals.

## Verbosity

data Verbosity Source

The verbosity data type, used by shakeVerbosity.

Constructors

 Silent Don't print any messages. Quiet Only print essential messages (typically errors). Normal Print normal messages (typically errors and warnings). Loud Print lots of messages (typically errors, warnings and status updates). Diagnostic Print messages for virtually everything (for debugging a build system).

Get the current verbosity level, as set by shakeVerbosity. If you want to output information to the console, you are recommended to use putLoud / putNormal / putQuiet, which ensures multiple messages are not interleaved.

Write a message to the output when the verbosity (shakeVerbosity) is appropriate. The output will not be interleaved with any other Shake messages (other than those generated by system commands).

Write a message to the output when the verbosity (shakeVerbosity) is appropriate. The output will not be interleaved with any other Shake messages (other than those generated by system commands).

Write a message to the output when the verbosity (shakeVerbosity) is appropriate. The output will not be interleaved with any other Shake messages (other than those generated by system commands).

quietly :: Action a -> Action aSource

Run an action with Quiet verbosity, in particular messages produced by traced (including from system') will not be printed to the screen.

# Utility functions

system' :: FilePath -> [String] -> Action ()Source

Execute a system command. This function will raise an error if the exit code is non-zero. Before running system' make sure you need any required files.

systemCwd :: FilePath -> FilePath -> [String] -> Action ()Source

Execute a system command with a specified current working directory (first argument). This function will raise an error if the exit code is non-zero. Before running systemCwd make sure you need any required files.

 systemCwd "/usr/MyDirectory" "pwd" []


systemOutput :: FilePath -> [String] -> Action (String, String)Source

Execute a system command, returning (stdout,stderr). This function will raise an error if the exit code is non-zero. Before running systemOutput make sure you need any required files.

copyFile old new copies the existing file from old to new. The old file is has need called on it before copying the file.

Read a file, after calling need.

A version of readFile' which also splits the result into lines.

Write a file, lifted to the Action monad.

writeFileLines :: FilePath -> [String] -> Action ()Source

A version of writeFile' which writes out a list of lines.

Write a file, but only if the contents would change.

# File rules

need :: [FilePath] -> Action ()Source

Require that the following files are built before continuing. Particularly necessary when calling system'. As an example:

 "*.rot13" *> \out -> do
let src = dropExtension out
need [src]
system' ["rot13",src,"-o",out]


want :: [FilePath] -> Rules ()Source

Require that the following are built by the rules, used to specify the target.

 main = shake shakeOptions $do want ["Main.exe"] ...  This program will build Main.exe, given sufficient rules. (*>) :: FilePattern -> (FilePath -> Action ()) -> Rules ()Source Define a rule that matches a FilePattern. No file required by the system must be matched by more than one pattern. For the pattern rules, see ?==.  "*.asm.o" *> \out -> do let src = dropExtension out need [src] system' ["as",src,"-o",out]  To define a build system for multiple compiled languages, we recommend using .asm.o, .cpp.o, .hs.o, to indicate which language produces an object file. I.e., the file foo.cpp produces object file foo.cpp.o. Note that matching is case-sensitive, even on Windows. (**>) :: [FilePattern] -> (FilePath -> Action ()) -> Rules ()Source Define a set of patterns, and if any of them match, run the associated rule. See *>. (?>) :: (FilePath -> Bool) -> (FilePath -> Action ()) -> Rules ()Source Define a rule to build files. If the first argument returns True for a given file, the second argument will be used to build it. Usually *> is sufficient, but ?> gives additional power. For any file used by the build system, only one rule should return True.  (all isUpper . takeBaseName) ?> \out -> do let src = replaceBaseName out$ map toLower $takeBaseName out writeFile' . map toUpper =<< readFile' src  (?>>) :: (FilePath -> Maybe [FilePath]) -> ([FilePath] -> Action ()) -> Rules ()Source Define a rule for building multiple files at the same time, a more powerful and more dangerous version of *>>. Given an application test ?>> ..., test should return Just if the rule applies, and should return the list of files that will be produced. This list must include the file passed as an argument and should obey the invariant: test x == Just ys ==> x elem ys && all ((== Just ys) . test) ys  As an example of a function satisfying the invariaint: test x | takeExtension x elem [".hi",".o"] = Just [dropExtension x <.> "hi", dropExtension x <.> "o"] test _ = Nothing  Regardless of whether Foo.hi or Foo.o is passed, the function always returns [Foo.hi, Foo.o]. (*>>) :: [FilePattern] -> ([FilePath] -> Action ()) -> Rules ()Source Define a rule for building multiple files at the same time. As an example, a single invokation of GHC produces both .hi and .o files:  ["*.o","*.hi"] *>> \[o,hi] -> do let hs = replaceExtension o "hs" need ... -- all files the .hs import system' "ghc" ["-c",hs]  However, in practice, it's usually easier to define rules with *> and make the .hi depend on the .o. When defining rules that build multiple files, all the FilePattern values must have the same sequence of // and * wildcards in the same order. A type synonym for file patterns, containing // and *. For the syntax and semantics of FilePattern see ?==. Match a FilePattern against a FilePath, There are only two special forms: • * matches an entire path component, excluding any separators. • // matches an arbitrary number of path components. Some examples that match:  "*.c" ?== "foo/bar/baz.c" "*.c" ?== "baz.c" "*.c" ?== "baz.c" "test.c" ?== "test.c"  Examples that don't match:  "*.c" ?== "foo/bar.c" "*/*.c" ?== "foo/bar/baz.c"  An example that only matches on Windows:  "foo/bar" ?== "foo\\bar"  # Directory rules Returns True if the file exists. Returns True if the directory exists. Get the contents of a directory. The result will be sorted, and will not contain the entries . or .. (unlike the standard Haskell version). The resulting paths will be relative to the first argument. It is usually simpler to call either getDirectoryFiles or getDirectoryDirs. Get the files anywhere under a directory that match any of a set of patterns. For the interpretation of the patterns see ?==. All results will be relative to the FilePath argument. Some examples:  getDirectoryFiles "Config" ["//*.xml"] -- All .xml files anywhere under the Config directory -- If Config/foo/bar.xml exists it will return ["foo/bar.xml"] getDirectoryFiles "Modules" ["*.hs","*.lhs"] -- All .hs or .lhs in the Modules directory -- If Modules/foo.hs and Modules/foo.lhs exist, it will return ["foo.hs","foo.lhs"]  Get the directories in a directory, not including . or ... All directories are relative to the argument directory.  getDirectoryDirs "/Users" -- Return all directories in the /Users directory -- e.g. ["Emily","Henry","Neil"]  # Environment rules Return Just the value of the environment variable, or Nothing if the variable is not set. # Oracle rules addOracle :: (ShakeValue q, ShakeValue a) => (q -> Action a) -> Rules (q -> Action a)Source Add extra information which rules can depend on. An oracle is a function from a question type q, to an answer type a. As an example, we can define an oracle allowing you to depend on the current version of GHC:  newtype GhcVersion = GhcVersion () deriving (Show,Typeable,Eq,Hashable,Binary,NFData) addOracle$ \(GhcVersion _) -> fmap (last . words . fst) $systemOutput "ghc" ["--version"]  If a rule calls askOracle (GhcVersion ()), that rule will be rerun whenever the GHC version changes. Some notes: • We define GhcVersion with a newtype around (), allowing the use of GeneralizedNewtypeDeriving. All the necessary type classes are exported from Development.Shake.Classes. • Each call to addOracle must use a different type of question. • Actions passed to addOracle will be run in every build they are required, but if their value does not change they will not invalidate any rules depending on them. To get a similar behaviour using data stored in files, see alwaysRerun. • If the value returned by askOracle is ignored then askOracleWith may help avoid ambiguous type messages. Alternatively, use the result of addOracle, which is askOracle restricted to the correct type. As a more complex example, consider tracking Haskell package versions: newtype GhcPkgList = GhcPkgList () deriving (Show,Typeable,Eq,Hashable,Binary,NFData) newtype GhcPkgVersion = GhcPkgVersion String deriving (Show,Typeable,Eq,Hashable,Binary,NFData) do getPkgList <- addOracle$ \GhcPkgList{} -> do
(out,_) <- systemOutput "ghc-pkg" ["list","--simple-output"]
return [(reverse b, reverse a) | x <- words out, let (a,_:b) = break (== '-') $reverse x] -- getPkgVersion <- addOracle$ \(GhcPkgVersion pkg) -> do
pkgs <- getPkgList
return $lookup pkg pkgs  Using these definitions, any rule depending on the version of shake should call getPkgVersion "shake" to rebuild when shake is upgraded. askOracle :: (ShakeValue q, ShakeValue a) => q -> Action aSource Get information previously added with addOracle. The question/answer types must match those provided to addOracle. askOracleWith :: (ShakeValue q, ShakeValue a) => q -> a -> Action aSource Get information previously added with addOracle. The second argument is not used, but can be useful to fix the answer type, avoiding ambiguous type error messages. # Special rules Always rerun the associated action. Useful for defining rules that query the environment. For example:  "ghcVersion.txt" *> \out -> do alwaysRerun (stdout,_) <- systemOutput "ghc" ["--version"] writeFileChanged out stdout  # Finite resources data Resource Source A type representing a finite resource, which multiple build actions should respect. Created with newResource and used with withResource when defining rules. As an example, only one set of calls to the Excel API can occur at one time, therefore Excel is a finite resource of quantity 1. You can write:  shake shakeOptions{shakeThreads=2}$ do
want ["a.xls","b.xls"]
excel <- newResource "Excel" 1
"*.xls" *> \out ->
withResource excel 1 $system' "excel" [out,...]  Now the two calls to excel will not happen in parallel. Using Resource is better than MVar as it will not block any other threads from executing. Be careful that the actions run within withResource do not themselves require further quantities of this resource, or you may get a "thread blocked indefinitely in an MVar operation" exception. Typically only system commands (such as system') will be run inside withResource, not commands such as need. As another example, calls to compilers are usually CPU bound but calls to linkers are usually disk bound. Running 8 linkers will often cause an 8 CPU system to grid to a halt. We can limit ourselves to 4 linkers with:  disk <- newResource "Disk" 4 want [show i <.> "exe" | i <- [1..100]] "*.exe" *> \out -> withResource disk 1$
system' "ld" ["-o",out,...]
"*.o" *> \out ->
system' "cl" ["-o",out,...]


Instances

 Show Resource

Create a new finite resource, given a name (for error messages) and a quantity of the resource that exists. For an example see Resource.

A version of newResource that runs in IO, and can be called before calling shake. Most people should use newResource instead.

withResource :: Resource -> Int -> Action a -> Action aSource

Run an action which uses part of a finite resource. For an example see Resource.

# Cached file contents

newCache :: (FilePath -> IO a) -> Rules (FilePath -> Action a)Source

Given a way of loading information from a file, produce a cached version that will load each file at most once. Using the cached function will still result in a dependency on the original file. The argument function should not access any files other than the one passed as its argument.

This function is useful when creating files that store intermediate values, to avoid the overhead of repeatedly reading from disk, particularly if the file requires expensive parsing.

newCacheIO :: (FilePath -> IO a) -> IO (FilePath -> Action a)Source

A version of newCache that runs in IO, and can be called before calling shake. Most people should use newCache instead.