add --cache-success snapshot flag

.gitignore

@@ -2,3 +2,4 @@
 *~
 .test-env
 ghtest-env
+.aider*

README.md

@@ -2,6 +2,7 @@
 
 ## Features
 
+
 - Runs tasks (which are defined as shell scripts)
 - Ensures only one instance of a task runs at a time (globally in the whole system) - using file-based locks
 - Tasks can run in parallel
@@ -128,3 +129,16 @@ To use it, first build using another system, and the run `taskrunner` with `TASK
 
 - Previous impl no-op tests/scripts/UPDATE: ~1.6s
 - Current impl no-op tests/scripts/UPDATE: ~2.3s
+
+## Snapshot Command Flags
+
+The `snapshot` command supports the following flags:
+
+- `--outputs`: Specifies the output files of the task. These files are used to determine if the task needs to be rerun.
+- `--cache-success`: Use remote cache even when no outputs are specified. The task is not rerun if it succeeded previously with the same inputs. Useful e.g. for test suites.
+- `--raw`: Specifies raw input strings that are used to compute the task's hash.
+- `--fuzzy-cache`: Enables the use of a fuzzy cache, which attempts to restore from a cache of a similar task if the exact cache is not available.
+- `--cache-root`: Specifies the root directory for caching. Use when caching things outside of the repository, e.g. `~/.stack`.
+- `--cache-version`: Specifies a version string for the cache. `--fuzzy-cache` will not download cache from another version, allowing clean breaks when making big changes, e.g. upgrading a compiler.
+- `--commit-status`: Enables reporting of the task's status to a commit status system, such as GitHub checks.
+- `--long-running`: Indicates that the task is expected to run for a long time (e.g. a server). Currently doens't have any effect though, TODO: can we remove it?

src/App.hs

@@ -197,7 +197,7 @@ main = do
           forM_ snapshotArgs.postUnpackCommands \cmd -> do
             runPostUnpackCmd appState cmd
 
-          when (hasOutputs snapshotArgs && settings.saveRemoteCache && (not skipped || appState.settings.primeCacheMode)) do
+          when (hasRemoteCache snapshotArgs && settings.saveRemoteCache && (not skipped || appState.settings.primeCacheMode)) do
             logDebug appState "Saving remote cache"
             s <- RemoteCache.getRemoteCacheSettingsFromEnv
             RemoteCache.saveCache appState s (fromMaybe settings.rootDirectory snapshotArgs.cacheRoot) snapshotArgs.outputs (archiveName appState snapshotArgs h.hash)
@@ -397,8 +397,8 @@ commandHandler appState requestPipe responsePipe =
 hasInputs :: SnapshotCliArgs -> Bool
 hasInputs args = not (null args.fileInputs) || not (null args.rawInputs)
 
-hasOutputs :: SnapshotCliArgs -> Bool
-hasOutputs args = not (null args.outputs)
+hasRemoteCache :: SnapshotCliArgs -> Bool
+hasRemoteCache args = args.cacheSuccess || not (null args.outputs)
 
 snapshot :: AppState -> SnapshotCliArgs -> IO String
 snapshot appState args = do
@@ -447,7 +447,7 @@ snapshot appState args = do
       earlyReturn (False, Nothing)
 
 
-    when (hasOutputs args && not force) do
+    when (hasRemoteCache args && not force) do
       s <- RemoteCache.getRemoteCacheSettingsFromEnv
       success <- liftIO $ RemoteCache.restoreCache appState s (fromMaybe appState.settings.rootDirectory args.cacheRoot) (archiveName appState args currentHash) RemoteCache.Log
       when success do
@@ -456,7 +456,7 @@ snapshot appState args = do
 
     logInfo appState "Inputs changed, running task"
 
-    when (not force && hasOutputs args && args.fuzzyCache && mainBranchCommitChanged savedHashInfo hashInfo) do
+    when (not force && hasRemoteCache args && args.fuzzyCache && mainBranchCommitChanged savedHashInfo hashInfo) do
       success <- tryRestoreFuzzyCache appState args
       when success do
         -- Save change in mainBranchCommit, even if the task didn't succeed yet.

src/RemoteCache.hs

@@ -40,7 +40,7 @@ import Amazonka.S3.PutObject (newPutObject, PutObject(..))
 packTar :: MonadResource m => AppState -> FilePath -> [FilePath] -> ConduitT () BS.ByteString m ()
 packTar appState workdir files = do
   let cmd = "tar"
-  let args = ["-c"] <> files
+  let args = if null files then ["-c", "--files-from=/dev/null"] else ["-c"] <> files
   liftIO $ logDebug appState $ "Running subprocess: " <> show (cmd:args) <> " in cwd " <> show workdir
   bracketP ( createProcess_ "createProcess_"
     (proc cmd args)

src/SnapshotCliArgs.hs

@@ -14,6 +14,7 @@ data SnapshotCliArgs = SnapshotCliArgs
     , cacheRoot :: Maybe FilePath
     , cacheVersion :: Maybe Text
     , commitStatus :: Bool
+    , cacheSuccess :: Bool
     } deriving (Show)
 
 instance Default SnapshotCliArgs where
@@ -28,6 +29,7 @@ instance Default SnapshotCliArgs where
     , cacheRoot = Nothing
     , cacheVersion = Nothing
     , commitStatus = False
+    , cacheSuccess = False
     }
 
 type ParseError = String
@@ -79,6 +81,10 @@ parse input = map fst $ flip execStateT (def :: SnapshotCliArgs, BeforeOutputs)
     modifyArgs (\s -> s { commitStatus = True })
     go xs
 
+  go ("--cache-success":xs) = do
+    modifyArgs (\s -> s { cacheSuccess = True })
+    go xs
+
   go (opt:_) | '-':_ <- opt =
     lift $ Left $ "Invalid option or missing argument: " <> opt
 

test/t/cache-success-changed.out

@@ -0,0 +1,9 @@
+-- output:
+[mytask] info | Inputs changed, running task
+[mytask] stdout | Expensive computation
+[mytask] stdout | oof
+[mytask] info | success
+[mytask] info | Inputs changed, running task
+[mytask] stdout | Expensive computation
+[mytask] stdout | rab
+[mytask] info | success

test/t/cache-success-changed.txt

@@ -0,0 +1,35 @@
+# no toplevel
+# s3
+
+export TASKRUNNER_SAVE_REMOTE_CACHE=1
+
+mkdir a
+
+(
+  cd a
+  echo foo > input.txt
+  git init -q
+  git add input.txt
+)
+
+cp -r a b
+
+go() {
+  # Note: we want separate workdir for separate repos
+  TASKRUNNER_STATE_DIRECTORY="$(pwd)" taskrunner -n mytask bash -e -c '
+    snapshot input.txt --cache-success
+    echo "Expensive computation"
+    rev input.txt
+  '
+}
+
+(
+  cd a
+  go
+)
+
+(
+  cd b
+  echo bar > input.txt
+  go # note: should run "expensive computation" again
+)

test/t/cache-success-unchanged.out

@@ -0,0 +1,7 @@
+-- output:
+[mytask] info | Inputs changed, running task
+[mytask] stdout | Expensive computation
+[mytask] stdout | oof
+[mytask] info | success
+[mytask] info | Found remote cache mytask-6b36308659163b577fd1c832107dc46ca3aa659b.tar.zst, restoring
+[mytask] info | Restored from remote cache

test/t/cache-success-unchanged.txt

@@ -0,0 +1,34 @@
+# no toplevel
+# s3
+
+export TASKRUNNER_SAVE_REMOTE_CACHE=1
+
+mkdir a
+
+(
+  cd a
+  echo foo > input.txt
+  git init -q
+  git add input.txt
+)
+
+cp -r a b
+
+go() {
+  # Note: we want separate workdir for separate repos
+  TASKRUNNER_STATE_DIRECTORY="$(pwd)" taskrunner -n mytask bash -e -c '
+    snapshot input.txt --cache-success
+    echo "Expensive computation"
+    rev input.txt
+  '
+}
+
+(
+  cd a
+  go
+)
+
+(
+  cd b
+  go # note: should not run "expensive computation" again
+)