litsupport/testfile.py

"""Parser for .test files"""
from lit.TestRunner import (
    parseIntegratedTestScriptCommands,
    getDefaultSubstitutions,
    applySubstitutions,
)
from litsupport import shellcommand
import logging

import os
import shlex

def _parseShellCommand(script, ln):
    # Trim trailing whitespace.
    ln = ln.rstrip()

    # Collapse lines with trailing '\\'.
    if script and script[-1][-1] == "\\":
        script[-1] = script[-1][:-1] + ln
    else:
        script.append(ln)

def clean_verify_command_for_windows(command_str):
    """Prepares a command string for execution in Windows command line by
    formatting paths and arguments.""" 
    # Replace backslashes with forward slashes and split into tokens
    tokens = shlex.split(command_str.replace("\\", "/"))
    # Enclose each token in double quotes and convert slashes back
    tokens_quoted = ['"{}"'.format(token.replace('/', '\\')) for token in tokens]
    # Join the tokens into a single command string and return
    return ' '.join(tokens_quoted)

def parse(context, filename):
    """Parse a .test file as used in the llvm test-suite.
    The file comporises of a number of lines starting with RUN: and VERIFY:
    specifying shell commands to run the benchmark and verifying the result.
    Returns a tuple with two arrays for the run and verify commands."""
    # Collect the test lines from the script.
    preparescript = []
    runscript = []
    verifyscript = []
    metricscripts = {}
    # Note that we keep both "RUN" and "RUN:" in the list to stay compatible
    # with older lit versions.
    keywords = [
        "PREPARE:",
        "PREPARE",
        "RUN:",
        "RUN",
        "VERIFY:",
        "VERIFY",
        "METRIC:",
        "METRIC",
    ]
    for line_number, command_type, ln in parseIntegratedTestScriptCommands(
        filename, keywords
    ):
        if command_type.startswith("PREPARE"):
            _parseShellCommand(preparescript, ln)
        elif command_type.startswith("RUN"):
            _parseShellCommand(runscript, ln)
        elif command_type.startswith("VERIFY"):
            _parseShellCommand(verifyscript, ln)
        elif command_type.startswith("METRIC"):
            metric, ln = ln.split(":", 1)
            metricscript = metricscripts.setdefault(metric.strip(), list())
            _parseShellCommand(metricscript, ln)
        else:
            raise ValueError("unknown script command type: %r" % (command_type,))

    # Verify the script contains a run line.
    if runscript == []:
        raise ValueError("Test has no RUN: line!")

    # Check for unterminated run lines.
    for script in preparescript, runscript, verifyscript:
        if script and script[-1][-1] == "\\":
            raise ValueError(
                "Test has unterminated RUN/VERIFY lines " + "(ending with '\\')"
            )

    # Apply the usual lit substitutions (%s, %S, %p, %T, ...)
    outfile = context.tmpBase + ".out"
    substitutions = getDefaultSubstitutions(
        context.test, context.tmpDir, context.tmpBase
    )
    substitutions += [("%o", outfile)]
    preparescript = applySubstitutions(preparescript, substitutions)
    runscript = applySubstitutions(runscript, substitutions)
    verifyscript = applySubstitutions(verifyscript, substitutions)
    if os.name == 'nt':
        verifyscript = [clean_verify_command_for_windows(cmd) for cmd in verifyscript]

    metricscripts = {
        k: applySubstitutions(v, substitutions) for k, v in metricscripts.items()
    }

    # Put things into the context
    context.parsed_preparescript = preparescript
    context.parsed_runscript = runscript
    context.parsed_verifyscript = verifyscript
    context.parsed_metricscripts = metricscripts
    context.executable = shellcommand.getMainExecutable(context)
    if not context.executable:
        logging.error("Could not determine executable name in %s" % filename)