[lit] Add support for readfile to external shell

llvm/docs/CommandGuide/lit.rst

@@ -664,6 +664,7 @@ TestRunner.py:
                          Otherwise, %t but with a single leading ``/`` removed.
  %:T                     On Windows, %/T but a ``:`` is removed if its the second character.
                          Otherwise, %T but with a single leading ``/`` removed.
+ %{readfile:<filename>}  Reads the file specified.
  ======================= ==============
 
 Other substitutions are provided that are variations on this base set and

llvm/utils/lit/lit/TestRunner.py

@@ -720,6 +720,26 @@ def processRedirects(cmd, stdin_source, cmd_shenv, opened_files):
     return std_fds
 
 
+def _expandLateSubstitutions(cmd, arguments, cwd):
+    for i, arg in enumerate(arguments):
+        if not isinstance(arg, str):
+            continue
+
+        def _replaceReadFile(match):
+            filePath = match.group(1)
+            if not os.path.isabs(filePath):
+                filePath = os.path.join(cwd, filePath)
+            try:
+                with open(filePath) as fileHandle:
+                    return fileHandle.read()
+            except FileNotFoundError:
+                raise InternalShellError(cmd, "File does not exist: %s" % filePath)
+
+        arguments[i] = re.sub(r"%{readfile:([^}]*)}", _replaceReadFile, arg)
+
+    return arguments
+
+
 def _executeShCmd(cmd, shenv, results, timeoutHelper):
     if timeoutHelper.timeoutReached():
         # Prevent further recursion if the timeout has been hit
@@ -834,6 +854,9 @@ def _executeShCmd(cmd, shenv, results, timeoutHelper):
         # Ensure args[0] is hashable.
         args[0] = expand_glob(args[0], cmd_shenv.cwd)[0]
 
+        # Expand all late substitutions.
+        args = _expandLateSubstitutions(j, args, cmd_shenv.cwd)
+
         inproc_builtin = inproc_builtins.get(args[0], None)
         if inproc_builtin and (args[0] != "echo" or len(cmd.commands) == 1):
             # env calling an in-process builtin is useless, so we take the safe
@@ -2389,6 +2412,20 @@ def runOnce(
         status, output, attempts=i + 1, max_allowed_attempts=attempts
     )
 
+def _expandLateSubstitutionsExternal(commandLine):
+    filePaths = []
+    def _replaceReadFile(match):
+        filePath = match.group(1)
+        filePaths.append(filePath)
+        return "$(cat %s)" % filePath
+
+    commandLine = re.sub(r"%{readfile:([^}]*)}", _replaceReadFile, commandLine)
+    # Add test commands before the command to check if the file exists as
+    # cat inside a subshell will never return a non-zero exit code outside
+    # of the subshell.
+    for filePath in filePaths:
+        commandLine = "%s && test -e %s" % (commandLine, filePath)
+    return commandLine
 
 def executeShTest(
     test, litConfig, useExternalSh, extra_substitutions=[], preamble_commands=[]
@@ -2420,4 +2457,8 @@ def executeShTest(
         recursion_limit=test.config.recursiveExpansionLimit,
     )
 
+    if useExternalSh:
+        for index, command in enumerate(script):
+            script[index] = _expandLateSubstitutionsExternal(command)
+
     return _runShTest(test, litConfig, useExternalSh, script, tmpBase)

llvm/utils/lit/tests/Inputs/shtest-readfile/absolute-paths.txt

@@ -0,0 +1,6 @@
+## Tests that readfile works with absolute paths.
+# RUN: echo -n "hello" > %t
+# RUN: echo %{readfile:%t}
+
+## Fail the test so we can assert on the output.
+# RUN: not echo return

llvm/utils/lit/tests/Inputs/shtest-readfile/file-does-not-exist.txt

@@ -0,0 +1,4 @@
+## Test that readfile reports information appropriately when the file specified
+## does not exist.
+
+# RUN: echo %{readfile:/file/does/not/exist}

llvm/utils/lit/tests/Inputs/shtest-readfile/lit.cfg

@@ -0,0 +1,19 @@
+import os
+
+import lit.formats
+import lit.util
+
+config.name = "shtest-readfile"
+config.suffixes = [".txt"]
+lit_shell_env = os.environ.get("LIT_USE_INTERNAL_SHELL")
+use_lit_shell = lit.util.pythonize_bool(lit_shell_env)
+config.test_format = lit.formats.ShTest(execute_external=not use_lit_shell)
+config.test_source_root = None
+config.test_exec_root = None
+
+# If we are testing with the external shell, remove the fake-externals from
+# PATH so that we use mkdir in the tests.
+if not use_lit_shell:
+    path_parts = config.environment["PATH"].split(os.path.pathsep)
+    path_parts = [path_part for path_part in path_parts if "fake-externals" not in path_part]
+    config.environment["PATH"] = os.path.pathsep.join(path_parts)

llvm/utils/lit/tests/Inputs/shtest-readfile/relative-paths.txt

@@ -0,0 +1,7 @@
+## Tests that readfile works with relative paths.
+# RUN: mkdir -p rel_path_test_folder
+# RUN: echo -n "hello" > rel_path_test_folder/test_file
+# RUN: echo %{readfile:rel_path_test_folder/test_file}
+
+## Fail the test so we can assert on the output.
+# RUN: not echo return

llvm/utils/lit/tests/Inputs/shtest-readfile/two-same-line.txt

@@ -0,0 +1,8 @@
+## Tests that readfile works with two substitutions on the same line to ensure the
+## regular expressions are set up correctly.
+# RUN: echo -n "hello" > %t.1
+# RUN: echo -n "bye" > %t.2
+# RUN: echo %{readfile:%t.1} %{readfile:%t.2}
+
+## Fail the test so we can assert on the output.
+# RUN: not echo return

llvm/utils/lit/tests/shtest-readfile-external.py

@@ -0,0 +1,21 @@
+## Tests the readfile substitution.
+
+# RUN: env LIT_USE_INTERNAL_SHELL=0 not %{lit} -a -v %{inputs}/shtest-readfile | FileCheck -match-full-lines -DTEMP_PATH=%S/Inputs/shtest-readfile/Output %s
+
+# CHECK: -- Testing: 4 tests{{.*}}
+
+# CHECK-LABEL: FAIL: shtest-readfile :: absolute-paths.txt ({{[^)]*}})
+# CHECK: echo $(cat [[TEMP_PATH]]/absolute-paths.txt.tmp) && test -e /home/gha/llvm-project/build/utils/lit/tests/Inputs/shtest-readfile/Output/absolute-paths.txt.tmp {{.*}}
+# CHECK: + echo hello
+
+# CHECK-LABEL: FAIL: shtest-readfile :: file-does-not-exist.txt ({{[^)]*}})
+# CHECK: echo $(cat /file/does/not/exist) && test -e /file/does/not/exist {{.*}}
+# CHECK: cat: /file/does/not/exist: No such file or directory
+
+# CHECK-LABEL: FAIL: shtest-readfile :: relative-paths.txt ({{[^)]*}})
+# CHECK: echo $(cat rel_path_test_folder/test_file) && test -e rel_path_test_folder/test_file {{.*}}
+# CHECK: + echo hello
+
+# CHECK-LABEL: FAIL: shtest-readfile :: two-same-line.txt ({{[^)]*}})
+# CHECK: echo $(cat /home/gha/llvm-project/build/utils/lit/tests/Inputs/shtest-readfile/Output/two-same-line.txt.tmp.1) $(cat /home/gha/llvm-project/build/utils/lit/tests/Inputs/shtest-readfile/Output/two-same-line.txt.tmp.2) && test -e /home/gha/llvm-project/build/utils/lit/tests/Inputs/shtest-readfile/Output/two-same-line.txt.tmp.1 && test -e /home/gha/llvm-project/build/utils/lit/tests/Inputs/shtest-readfile/Output/two-same-line.txt.tmp.2 {{.*}}
+# CHECK: + echo hello bye

llvm/utils/lit/tests/shtest-readfile.py

@@ -0,0 +1,21 @@
+## Tests the readfile substitution.
+
+# RUN: env LIT_USE_INTERNAL_SHELL=1 not %{lit} -a -v %{inputs}/shtest-readfile | FileCheck -match-full-lines -DTEMP_PATH=%S/Inputs/shtest-readfile/Output %s
+
+# CHECK: -- Testing: 4 tests{{.*}}
+
+# CHECK-LABEL: FAIL: shtest-readfile :: absolute-paths.txt ({{[^)]*}})
+# CHECK: echo hello
+# CHECK: # executed command: echo '%{readfile:[[TEMP_PATH]]/absolute-paths.txt.tmp}'
+
+# CHECK-LABEL: FAIL: shtest-readfile :: file-does-not-exist.txt ({{[^)]*}})
+# CHECK: # executed command: @echo 'echo %{readfile:/file/does/not/exist}'
+# CHECK: # | File does not exist: /file/does/not/exist
+
+# CHECK-LABEL: FAIL: shtest-readfile :: relative-paths.txt ({{[^)]*}})
+# CHECK: echo hello
+# CHECK: # executed command: echo '%{readfile:rel_path_test_folder/test_file}'
+
+# CHECK-LABEL: FAIL: shtest-readfile :: two-same-line.txt ({{[^)]*}})
+# CHECK: echo hello bye
+# CHECK: # executed command: echo '%{readfile:[[TEMP_PATH]]/two-same-line.txt.tmp.1}' '%{readfile:[[TEMP_PATH]]/two-same-line.txt.tmp.2}'