verilator/nodist/fastcov.py

#!/usr/bin/env python3
# SPDX-License-Identifier: MIT
# Copyright 2018-present, Bryan Gillespie
"""
    Author: Bryan Gillespie
    https://github.com/RPGillespie6/fastcov

    A massively parallel gcov wrapper for generating intermediate coverage formats fast

    The goal of fastcov is to generate code coverage intermediate formats as fast as possible,
    even for large projects with hundreds of gcda objects. The intermediate formats may then be
    consumed by a report generator such as lcov's genhtml, or a dedicated frontend such as coveralls.

    Sample Usage:
        $ cd build_dir
        $ ./fastcov.py --zerocounters
        $ <run unit tests>
        $ ./fastcov.py --exclude /usr/include test/ --lcov -o report.info
        $ genhtml -o code_coverage report.info
"""
import re
import os
import sys
import glob
import json
import time
import fnmatch
import logging
import argparse
import threading
import subprocess
import multiprocessing
from pathlib import Path

FASTCOV_VERSION = (1,15)
MINIMUM_PYTHON  = (3,5)
MINIMUM_GCOV    = (9,0,0)

# Interesting metrics
START_TIME    = time.monotonic()
GCOVS_TOTAL   = 0
GCOVS_SKIPPED = 0

# Gcov Coverage File Extensions
GCOV_GCNO_EXT = ".gcno" # gcno = "[gc]ov [no]te"
GCOV_GCDA_EXT = ".gcda" # gcda = "[gc]ov [da]ta"

# For when things go wrong...
# Start error codes at 3 because 1-2 are special
# See https://stackoverflow.com/a/1535733/2516916
EXIT_CODE  = 0
EXIT_CODES = {
    "gcov_version": 3,
    "python_version": 4,
    "unsupported_coverage_format": 5,
    "excl_not_found": 6,
    "bad_chunk_file": 7,
    "missing_json_key": 8,
}

# Disable all logging in case developers are using this as a module
logging.disable(level=logging.CRITICAL)

class FastcovFormatter(logging.Formatter):
    def format(self, record):
        record.levelname = record.levelname.lower()
        log_message = super(FastcovFormatter, self).format(record)
        return "[{:.3f}s] {}".format(stopwatch(), log_message)

class DiffParseError(Exception):
    pass

class DiffParser(object):
    def _refinePaths(self, diff_metadata, diff_base_dir):
        diff_metadata.pop('/dev/null', None)
        diff_metadata.pop('', None)
        for key, value in diff_metadata.copy().items():
            diff_metadata.pop(key)
            #sources without added lines will be excluded
            if value:
                newpath = os.path.join(diff_base_dir, key) if diff_base_dir else os.path.abspath(key)
                diff_metadata[newpath] = value

    def _parseTargetFile(self, line_with_target_file):
        #f.e. '+++ b/README.md1' or '+++ b/README.md1   timestamp'
        target_source = line_with_target_file[4:].partition('\t')[0].strip()
        target_source = target_source[2:] if target_source.startswith('b/') else target_source
        return target_source

    def _parseHunkBoundaries(self, line_with_hunk_boundaries, line_index):
        #f.e. '@@ -121,4 +122,4 @@ Time to process all gcda and parse all gcov:'
        # Here ['-121,4', '+122,4']
        lines_info = line_with_hunk_boundaries[3:].partition("@@")[0].strip().split(' ')
        if len(lines_info) != 2:
            raise DiffParseError("Found invalid hunk. Line #{}. {}".format(line_index, line_with_hunk_boundaries))
        # Here ['122','4']
        target_lines_info = lines_info[1].strip('+').partition(',')
        target_line_current = int(target_lines_info[0])
        target_lines_count = int(target_lines_info[2]) if target_lines_info[2] else 1

        # Here ['121','4']
        source_lines_info = lines_info[0].strip('-').partition(',')
        source_line_current = int(source_lines_info[0])
        source_lines_count = int(source_lines_info[2]) if source_lines_info[2] else 1
        return target_line_current, target_lines_count, source_line_current, source_lines_count

    def parseDiffFile(self, diff_file, diff_base_dir, fallback_encodings=[]):
        diff_metadata = {}
        target_source = None
        target_hunk = set()
        target_line_current = 0
        target_line_end = 0
        source_line_current = 0
        source_line_end = 0
        found_hunk = False
        for i, line in enumerate(getSourceLines(diff_file, fallback_encodings), 1):
            line = line.rstrip()
            if not found_hunk:
                if line.startswith('+++ '):
                    # refresh file
                    target_source = self._parseTargetFile(line)
                elif line.startswith('@@ '):
                    # refresh hunk
                    target_line_current, target_lines_count, source_line_current, source_lines_count =  self._parseHunkBoundaries(line, i)
                    target_line_end = target_line_current + target_lines_count
                    source_line_end = source_line_current + source_lines_count
                    target_hunk = set()
                    found_hunk = True
                continue
            if target_line_current > target_line_end or source_line_current > source_line_end:
                raise DiffParseError("Hunk longer than expected. Line #{}. {}".format(i, line))

            if line.startswith('+'):
                #line related to target
                target_hunk.add(target_line_current)
                target_line_current = target_line_current + 1
            elif line.startswith(' ') or line == '':
                # line related to both
                target_line_current = target_line_current + 1
                source_line_current = source_line_current + 1
            elif line.startswith('-'):
                # line related to source
                source_line_current = source_line_current + 1
            elif not line.startswith('\\'):  # No newline at end of file
                # line with newline marker is not included into any boundaries
                raise DiffParseError("Found unrecognized hunk line type. Line #{}. {}".format(i, line))

            if target_line_current == target_line_end and source_line_current == source_line_end:
                # Checked all lines, save data
                if target_source in diff_metadata:
                    diff_metadata[target_source] = target_hunk.union(diff_metadata[target_source])
                else:
                    diff_metadata[target_source] = target_hunk
                target_hunk = set()
                found_hunk = False

        if target_line_current != target_line_end or source_line_current != source_line_end:
            raise DiffParseError("Unexpected end of file. Expected hunk with {} target lines, {} source lines".format(
                target_line_end - target_line_current, source_line_end - source_line_current))

        self._refinePaths(diff_metadata, diff_base_dir)
        return diff_metadata

    def filterByDiff(self, diff_file, dir_base_dir, fastcov_json, fallback_encodings=[]):
        diff_metadata = self.parseDiffFile(diff_file, dir_base_dir, fallback_encodings)
        logging.debug("Include only next files: {}".format(diff_metadata.keys()))
        excluded_files_count = 0
        excluded_lines_count = 0
        for source in list(fastcov_json["sources"].keys()):
            diff_lines = diff_metadata.get(source, None)
            if not diff_lines:
                excluded_files_count = excluded_files_count + 1
                logging.debug("Exclude {} according to diff file".format(source))
                fastcov_json["sources"].pop(source)
                continue
            for test_name, report_data in fastcov_json["sources"][source].copy().items():
                #No info about functions boundaries, removing all
                for function in list(report_data["functions"].keys()):
                    report_data["functions"].pop(function, None)
                for line in list(report_data["lines"].keys()):
                    if line not in diff_lines:
                        excluded_lines_count = excluded_lines_count + 1
                        report_data["lines"].pop(line)
                for branch_line in list(report_data["branches"].keys()):
                    if branch_line not in diff_lines:
                        report_data["branches"].pop(branch_line)
                if len(report_data["lines"]) == 0:
                    fastcov_json["sources"][source].pop(test_name)
            if len(fastcov_json["sources"][source]) == 0:
                excluded_files_count = excluded_files_count + 1
                logging.debug('Exclude {} file as it has no lines due to diff filter'.format(source))
                fastcov_json["sources"].pop(source)
        logging.info("Excluded {} files and {} lines according to diff file".format(excluded_files_count, excluded_lines_count))
        return fastcov_json

def chunks(l, n):
    """Yield successive n-sized chunks from l."""
    for i in range(0, len(l), n):
        yield l[i:i + n]

def setExitCode(key):
    global EXIT_CODE
    EXIT_CODE = EXIT_CODES[key]

def setExitCodeRaw(code):
    global EXIT_CODE
    EXIT_CODE = code

def incrementCounters(total, skipped):
    global GCOVS_TOTAL
    global GCOVS_SKIPPED
    GCOVS_TOTAL   += total
    GCOVS_SKIPPED += skipped

def stopwatch():
    """Return number of seconds since last time this was called."""
    global START_TIME
    end_time   = time.monotonic()
    delta      = end_time - START_TIME
    START_TIME = end_time
    return delta

def parseVersionFromLine(version_str):
    """Given a string containing a dotted integer version, parse out integers and return as tuple."""
    version = re.search(r'(\d+\.\d+\.\d+)', version_str)

    if not version:
        return (0,0,0)

    return tuple(map(int, version.group(1).split(".")))

def getGcovVersion(gcov):
    p = subprocess.Popen([gcov, "-v"], stdout=subprocess.PIPE)
    output = p.communicate()[0].decode('UTF-8')
    p.wait()
    return parseVersionFromLine(output.split("\n")[0])

def tryParseNumber(s):
    try:
        return int(s)
    except ValueError:
        # Log a warning if not hyphen
        if s != "-":
            logging.warning("Unsupported numerical value '%s', using 0", s)

        # Default to 0 if we can't parse the number (e.g. "-", "NaN", etc.)
        return 0

def removeFiles(files):
    for file in files:
        os.remove(file)

def processPrefix(path, prefix, prefix_strip):
    p = Path(path)
    if p.exists() or not p.is_absolute():
        return path

    if prefix_strip > 0:
        segments = p.parts

        if len(segments) < prefix_strip + 1:
            logging.warning("Couldn't strip %i path levels from %s.", prefix_strip, path)
            return path

        segments = segments[prefix_strip+1:]
        p = Path(segments[0])
        segments = segments[1:]
        for s in segments:
            p = p.joinpath(s)

    if len(prefix) > 0:
        if p.is_absolute():
            p = Path(prefix).joinpath(p.relative_to('/'))
        else:
            p = Path(prefix).joinpath(p)

    return str(p)


def getFilteredCoverageFiles(coverage_files, exclude):
    def excludeGcda(gcda):
        for ex in exclude:
            if ex in gcda:
                logging.debug("Omitting %s due to '--exclude-gcda %s'", gcda, ex)
                return False
        return True
    return list(filter(excludeGcda, coverage_files))

def globCoverageFiles(cwd, coverage_type):
    return glob.glob(os.path.join(os.path.abspath(cwd), "**/*" + coverage_type), recursive=True)

def findCoverageFiles(cwd, coverage_files, use_gcno):
    coverage_type = "user provided"
    if not coverage_files:
        # gcov strips off extension of whatever you pass it and searches [extensionless name] + .gcno/.gcda
        # We should pass either gcno or gcda, but not both - if you pass both it will be processed twice
        coverage_type = GCOV_GCNO_EXT if use_gcno else GCOV_GCDA_EXT
        coverage_files = globCoverageFiles(cwd, coverage_type)

    logging.info("Found {} coverage files ({})".format(len(coverage_files), coverage_type))
    logging.debug("Coverage files found:\n    %s", "\n    ".join(coverage_files))
    return coverage_files

def gcovWorker(data_q, metrics_q, args, chunk, gcov_filter_options):
    base_report   = {"sources": {}}
    gcovs_total   = 0
    gcovs_skipped = 0
    error_exit    = False

    gcov_bin = args.gcov
    gcov_args = ["--json-format", "--stdout"]
    if args.branchcoverage or args.xbranchcoverage:
        gcov_args.append("--branch-probabilities")

    encoding = sys.stdout.encoding if sys.stdout.encoding else 'UTF-8'
    workdir  = args.cdirectory if args.cdirectory else "."

    p = subprocess.Popen([gcov_bin] + gcov_args + chunk, cwd=workdir, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL)
    for i, line in enumerate(iter(p.stdout.readline, b'')):
        try:
            intermediate_json = json.loads(line.decode(encoding))
        except json.decoder.JSONDecodeError as e:
            logging.error("Could not process chunk file '{}' ({}/{})".format(chunk[i], i+1, len(chunk)))
            logging.error(str(e))
            setExitCode("bad_chunk_file")
            continue

        if "current_working_directory" not in intermediate_json:
            logging.error("Missing 'current_working_directory' for data file: {}".format(intermediate_json))
            setExitCode("missing_json_key")
            continue

        intermediate_json_files = processGcovs(args.cdirectory, intermediate_json["files"], intermediate_json["current_working_directory"], gcov_filter_options)
        for f in intermediate_json_files:
            distillSource(f, base_report["sources"], args.test_name, args.xbranchcoverage)
        gcovs_total   += len(intermediate_json["files"])
        gcovs_skipped += len(intermediate_json["files"]) - len(intermediate_json_files)

    p.wait()
    data_q.put(base_report)
    metrics_q.put((gcovs_total, gcovs_skipped))

    sys.exit(EXIT_CODE)

def processGcdas(args, coverage_files, gcov_filter_options):
    chunk_size = max(args.minimum_chunk, int(len(coverage_files) / args.jobs) + 1)

    processes = []
    data_q    = multiprocessing.Queue()
    metrics_q = multiprocessing.Queue()
    for chunk in chunks(coverage_files, chunk_size):
        p = multiprocessing.Process(target=gcovWorker, args=(data_q, metrics_q, args, chunk, gcov_filter_options))
        processes.append(p)
        p.start()

    logging.info("Spawned {} gcov processes, each processing at most {} coverage files".format(len(processes), chunk_size))

    fastcov_jsons = []
    for p in processes:
        fastcov_jsons.append(data_q.get())
        incrementCounters(*metrics_q.get())

    for p in processes:
        p.join()
        if p.exitcode != 0:
            setExitCodeRaw(p.exitcode)

    base_fastcov = fastcov_jsons.pop()
    for fj in fastcov_jsons:
        combineReports(base_fastcov, fj)

    return base_fastcov

def shouldFilterSource(source, gcov_filter_options):
    """Returns true if the provided source file should be filtered due to CLI options, otherwise returns false."""
    # If explicit sources were passed, check for match
    if gcov_filter_options["sources"]:
        if source not in gcov_filter_options["sources"]:
            logging.debug("Filtering coverage for '%s' due to option '--source-files'", source)
            return True

    # Check exclude filter
    for ex in gcov_filter_options["exclude"]:
        if ex in source:
            logging.debug("Filtering coverage for '%s' due to option '--exclude %s'", source, ex)
            return True

    # Check exclude filter
    for ex_glob in gcov_filter_options["exclude_glob"]:
        if fnmatch.fnmatch(source, ex_glob):
            logging.debug("Filtering coverage for '%s' due to option '--exclude-glob %s'", source, ex_glob)
            return True

    # Check include filter
    if gcov_filter_options["include"]:
        included = False
        for inc in gcov_filter_options["include"]:
            if inc in source:
                included = True
                break

        if not included:
            logging.debug("Filtering coverage for '%s' due to option '--include %s'", source, " ".join(gcov_filter_options["include"]))
            return True

    return False

def filterFastcov(fastcov_json, args):
    logging.info("Performing filtering operations (if applicable)")
    gcov_filter_options = getGcovFilterOptions(args)
    for source in list(fastcov_json["sources"].keys()):
        if shouldFilterSource(source, gcov_filter_options):
            del fastcov_json["sources"][source]

def processGcov(cwd, gcov, source_base_dir, files, gcov_filter_options):
    # Uses cwd if set, else source_base_dir from gcov json. If both are empty, uses "."
    base_dir = cwd if cwd else source_base_dir
    base_dir = base_dir if base_dir else "."
    # Add absolute path
    gcov["file_abs"] = os.path.abspath(os.path.join(base_dir, gcov["file"]))

    if shouldFilterSource(gcov["file_abs"], gcov_filter_options):
        return

    files.append(gcov)
    logging.debug("Accepted coverage for '%s'", gcov["file_abs"])

def processGcovs(cwd, gcov_files, source_base_dir, gcov_filter_options):
    files = []
    for gcov in gcov_files:
        processGcov(cwd, gcov, source_base_dir, files, gcov_filter_options)
    return files

def dumpBranchCoverageToLcovInfo(f, branches):
    branch_miss = 0
    branch_found = 0
    brda = []
    for line_num, branch_counts in branches.items():
        for i, count in enumerate(branch_counts):
            # Branch (<line number>, <block number>, <branch number>, <taken>)
            brda.append((line_num, int(i/2), i, count))
            branch_miss += int(count == 0)
            branch_found += 1
    for v in sorted(brda):
        f.write("BRDA:{},{},{},{}\n".format(*v))
    f.write("BRF:{}\n".format(branch_found))                # Branches Found
    f.write("BRH:{}\n".format(branch_found - branch_miss))  # Branches Hit

def dumpToLcovInfo(fastcov_json, output):
    with open(output, "w") as f:
        sources = fastcov_json["sources"]
        for sf in sorted(sources.keys()):
            for tn in sorted(sources[sf].keys()):
                data = sources[sf][tn]
                f.write("TN:{}\n".format(tn)) #Test Name - used mainly in conjuction with genhtml --show-details
                f.write("SF:{}\n".format(sf)) #Source File

                fn_miss = 0
                fn = []
                fnda = []
                for function, fdata in data["functions"].items():
                    fn.append((fdata["start_line"], function))  # Function Start Line
                    fnda.append((fdata["execution_count"], function))  # Function Hits
                    fn_miss += int(fdata["execution_count"] == 0)
                # NOTE: lcov sorts FN, but not FNDA.
                for v in sorted(fn):
                    f.write("FN:{},{}\n".format(*v))
                for v in sorted(fnda):
                    f.write("FNDA:{},{}\n".format(*v))
                f.write("FNF:{}\n".format(len(data["functions"])))               #Functions Found
                f.write("FNH:{}\n".format((len(data["functions"]) - fn_miss)))   #Functions Hit

                if data["branches"]:
                    dumpBranchCoverageToLcovInfo(f, data["branches"])

                line_miss = 0
                da = []
                for line_num, count in data["lines"].items():
                    da.append((line_num, count))
                    line_miss += int(count == 0)
                for v in sorted(da):
                    f.write("DA:{},{}\n".format(*v))  # Line
                f.write("LF:{}\n".format(len(data["lines"])))                 #Lines Found
                f.write("LH:{}\n".format((len(data["lines"]) - line_miss)))   #Lines Hit
                f.write("end_of_record\n")

def getSourceLines(source, fallback_encodings=[]):
    """Return a list of lines from the provided source, trying to decode with fallback encodings if the default fails."""
    default_encoding = sys.getdefaultencoding()
    for encoding in [default_encoding] + fallback_encodings:
        try:
            with open(source, encoding=encoding) as f:
                return f.readlines()
        except UnicodeDecodeError:
            pass

    logging.warning("Could not decode '{}' with {} or fallback encodings ({}); ignoring errors".format(source, default_encoding, ",".join(fallback_encodings)))
    with open(source, errors="ignore") as f:
        return f.readlines()

def containsMarker(markers, strBody):
    for marker in markers:
        if marker in strBody:
            return True
    return False

# Returns whether source coverage changed or not
def exclProcessSource(fastcov_sources, source, exclude_branches_sw, include_branches_sw, exclude_line_marker, fallback_encodings, gcov_prefix, gcov_prefix_strip):
    source_to_open = processPrefix(source, gcov_prefix, gcov_prefix_strip)

    # Before doing any work, check if this file even needs to be processed
    if not exclude_branches_sw and not include_branches_sw:
        # Ignore unencodable characters
        with open(source_to_open, errors="ignore") as f:
            if not containsMarker(exclude_line_marker + ["LCOV_EXCL"], f.read()):
                return False

    # If we've made it this far we have to check every line

    start_line = 0
    end_line = 0
    # Start enumeration at line 1 because the first line of the file is line 1 not 0
    for i, line in enumerate(getSourceLines(source_to_open, fallback_encodings), 1):
        # Cycle through test names (likely only 1)
        for test_name in fastcov_sources[source]:
            fastcov_data = fastcov_sources[source][test_name]

            # Check if branch coverage should be deleted based on CLI options
            if (exclude_branches_sw or include_branches_sw) and (i in fastcov_data["branches"]):
                del_exclude_br = exclude_branches_sw and any(line.lstrip().startswith(e) for e in exclude_branches_sw)
                del_include_br = include_branches_sw and all(not line.lstrip().startswith(e) for e in include_branches_sw)
                if del_exclude_br or del_include_br:
                    del fastcov_data["branches"][i]

            # Skip to next line as soon as possible
            if not containsMarker(exclude_line_marker + ["LCOV_EXCL"], line):
                continue

            # Build line to function dict so can quickly delete by line number
            line_to_func = {}
            for f in fastcov_data["functions"].keys():
                l = fastcov_data["functions"][f]["start_line"]
                if l not in line_to_func:
                    line_to_func[l] = set()
                line_to_func[l].add(f)

            if any(marker in line for marker in exclude_line_marker):
                for key in ["lines", "branches"]:
                    if i in fastcov_data[key]:
                        del fastcov_data[key][i]
                if i in line_to_func:
                    for key in line_to_func[i]:
                        if key in fastcov_data["functions"]:
                            del fastcov_data["functions"][key]
            elif "LCOV_EXCL_START" in line:
                start_line = i
            elif "LCOV_EXCL_STOP" in line:
                end_line = i

                if not start_line:
                    end_line = 0
                    continue

                for key in ["lines", "branches"]:
                    for line_num in list(fastcov_data[key].keys()):
                        if start_line <= line_num <= end_line:
                            del fastcov_data[key][line_num]

                for line_num in range(start_line, end_line):
                    if line_num in line_to_func:
                        for key in line_to_func[line_num]:
                            if key in fastcov_data["functions"]:
                                del fastcov_data["functions"][key]

                start_line = end_line = 0
            elif "LCOV_EXCL_BR_LINE" in line:
                if i in fastcov_data["branches"]:
                    del fastcov_data["branches"][i]

    # Source coverage changed
    return True

def exclMarkerWorker(data_q, fastcov_sources, chunk, exclude_branches_sw, include_branches_sw, exclude_line_marker, fallback_encodings, gcov_prefix, gcov_prefix_strip):
    changed_sources = []

    for source in chunk:
        try:
            if exclProcessSource(fastcov_sources, source, exclude_branches_sw, include_branches_sw, exclude_line_marker, fallback_encodings, gcov_prefix, gcov_prefix_strip):
                changed_sources.append((source, fastcov_sources[source]))
        except FileNotFoundError:
            logging.error("Could not find '%s' to scan for exclusion markers...", source)
            setExitCode("excl_not_found") # Set exit code because of error

    # Write out changed sources back to main fastcov file
    data_q.put(changed_sources)

    # Exit current process with appropriate code
    sys.exit(EXIT_CODE)

def processExclusionMarkers(fastcov_json, jobs, exclude_branches_sw, include_branches_sw, exclude_line_marker, min_chunk_size, fallback_encodings, gcov_prefix, gcov_prefix_strip):
    chunk_size = max(min_chunk_size, int(len(fastcov_json["sources"]) / jobs) + 1)

    processes = []
    data_q    = multiprocessing.Queue()
    for chunk in chunks(list(fastcov_json["sources"].keys()), chunk_size):
        p = multiprocessing.Process(target=exclMarkerWorker, args=(data_q, fastcov_json["sources"], chunk, exclude_branches_sw, include_branches_sw, exclude_line_marker, fallback_encodings, gcov_prefix, gcov_prefix_strip))
        processes.append(p)
        p.start()

    logging.info("Spawned {} exclusion marker scanning processes, each processing at most {} source files".format(len(processes), chunk_size))

    changed_sources = []
    for p in processes:
        changed_sources += data_q.get()

    for p in processes:
        p.join()
        if p.exitcode != 0:
            setExitCodeRaw(p.exitcode)

    for changed_source in changed_sources:
        fastcov_json["sources"][changed_source[0]] = changed_source[1]

def validateSources(fastcov_json, gcov_prefix, gcov_prefix_strip):
    logging.info("Checking if all sources exist")
    for source in fastcov_json["sources"].keys():
        source = processPrefix(source, gcov_prefix, gcov_prefix_strip)
        if not os.path.exists(source):
            logging.error("Cannot find '{}'".format(source))

def distillFunction(function_raw, functions):
    function_name   = function_raw["name"]
    # NOTE: need to explicitly cast all counts coming from gcov to int - this is because gcov's json library
    # will pass as scientific notation (i.e. 12+e45)
    start_line      = int(function_raw["start_line"])
    execution_count = int(function_raw["execution_count"])
    if function_name not in functions:
        functions[function_name] = {
            "start_line": start_line,
            "execution_count": execution_count
        }
    else:
        functions[function_name]["execution_count"] += execution_count

def emptyBranchSet(branch1, branch2):
    return (branch1["count"] == 0 and branch2["count"] == 0)

def matchingBranchSet(branch1, branch2):
    return (branch1["count"] == branch2["count"])

def filterExceptionalBranches(branches):
    filtered_branches = []
    exception_branch = False
    for i in range(0, len(branches), 2):
        if i+1 >= len(branches):
            filtered_branches.append(branches[i])
            break

        # Filter exceptional branch noise
        if branches[i+1]["throw"]:
            exception_branch = True
            continue

        # Filter initializer list noise
        if exception_branch and emptyBranchSet(branches[i], branches[i+1]) and len(filtered_branches) >= 2 and matchingBranchSet(filtered_branches[-1], filtered_branches[-2]):
            return []

        filtered_branches.append(branches[i])
        filtered_branches.append(branches[i+1])

    return filtered_branches

def distillLine(line_raw, lines, branches, include_exceptional_branches):
    line_number = int(line_raw["line_number"])
    count       = int(line_raw["count"])
    if count <  0:
        if "function_name" in line_raw:
            logging.warning("Ignoring negative count found in '%s'.", line_raw["function_name"])
        else:
            logging.warning("Ignoring negative count.")
        count = 0

    if line_number not in lines:
        lines[line_number] = count
    else:
        lines[line_number] += count

    # Filter out exceptional branches by default unless requested otherwise
    if not include_exceptional_branches:
        line_raw["branches"] = filterExceptionalBranches(line_raw["branches"])

    # Increment all branch counts
    for i, branch in enumerate(line_raw["branches"]):
        if line_number not in branches:
            branches[line_number] = []
        blen = len(branches[line_number])
        glen = len(line_raw["branches"])
        if blen < glen:
            branches[line_number] += [0] * (glen - blen)
        branches[line_number][i] += int(branch["count"])

def distillSource(source_raw, sources, test_name, include_exceptional_branches):
    source_name = source_raw["file_abs"]
    if source_name not in sources:
        sources[source_name] = {
            test_name: {
                "functions": {},
                "branches": {},
                "lines": {}
            }
        }

    for function in source_raw["functions"]:
        distillFunction(function, sources[source_name][test_name]["functions"])

    for line in source_raw["lines"]:
        distillLine(line, sources[source_name][test_name]["lines"], sources[source_name][test_name]["branches"], include_exceptional_branches)

def dumpToJson(intermediate, output):
    with open(output, "w") as f:
        json.dump(intermediate, f)

def getGcovFilterOptions(args):
    return {
        "sources": set([os.path.abspath(s) for s in args.sources]), #Make paths absolute, use set for fast lookups
        "include": args.includepost,
        "exclude": args.excludepost,
        "exclude_glob":args.excludepost_glob
    }

def addDicts(dict1, dict2):
    """Add dicts together by value. i.e. addDicts({"a":1,"b":0}, {"a":2}) == {"a":3,"b":0}."""
    result = {k:v for k,v in dict1.items()}
    for k,v in dict2.items():
        if k in result:
            result[k] += v
        else:
            result[k] = v

    return result

def addLists(list1, list2):
    """Add lists together by value. i.e. addLists([1,1], [2,2]) == [3,3]."""
    # Find big list and small list
    blist, slist = list(list2), list(list1)
    if len(list1) > len(list2):
        blist, slist = slist, blist

    # Overlay small list onto big list
    for i, b in enumerate(slist):
        blist[i] += b

    return blist

def combineReports(base, overlay):
    for source, scov in overlay["sources"].items():
        # Combine Source Coverage
        if source not in base["sources"]:
            base["sources"][source] = scov
            continue

        for test_name, tcov in scov.items():
            # Combine Source Test Name Coverage
            if test_name not in base["sources"][source]:
                base["sources"][source][test_name] = tcov
                continue

            # Drill down and create convenience variable
            base_data = base["sources"][source][test_name]

            # Combine Line Coverage
            base_data["lines"] = addDicts(base_data["lines"], tcov["lines"])

            # Combine Branch Coverage
            for branch, cov in tcov["branches"].items():
                if branch not in base_data["branches"]:
                    base_data["branches"][branch] = cov
                else:
                    base_data["branches"][branch] = addLists(base_data["branches"][branch], cov)

            # Combine Function Coverage
            for function, cov in tcov["functions"].items():
                if function not in base_data["functions"]:
                    base_data["functions"][function] = cov
                else:
                    base_data["functions"][function]["execution_count"] += cov["execution_count"]

def parseInfo(path):
    """Parse an lcov .info file into fastcov json."""
    fastcov_json = {
        "sources": {}
    }

    with open(path) as f:
        for line in f:
            if line.startswith("TN:"):
                current_test_name = line[3:].strip()
            elif line.startswith("SF:"):
                current_sf = line[3:].strip()
                fastcov_json["sources"].setdefault(current_sf, {
                    current_test_name: {
                        "functions": {},
                        "branches": {},
                        "lines": {},
                    }
                })
                current_data = fastcov_json["sources"][current_sf][current_test_name]
            elif line.startswith("FN:"):
                line_num, function_name = line[3:].strip().split(",")
                current_data["functions"][function_name] = {}
                current_data["functions"][function_name]["start_line"] = tryParseNumber(line_num)
            elif line.startswith("FNDA:"):
                count, function_name = line[5:].strip().split(",")
                current_data["functions"][function_name]["execution_count"] = tryParseNumber(count)
            elif line.startswith("DA:"):
                line_num, count = line[3:].strip().split(",")
                current_data["lines"][line_num] = tryParseNumber(count)
            elif line.startswith("BRDA:"):
                branch_tokens = line[5:].strip().split(",")
                line_num, count = branch_tokens[0], branch_tokens[-1]
                if line_num not in current_data["branches"]:
                    current_data["branches"][line_num] = []
                current_data["branches"][line_num].append(tryParseNumber(count))

    return fastcov_json

def convertKeysToInt(report):
    for source in report["sources"].keys():
        for test_name in report["sources"][source].keys():
            report_data = report["sources"][source][test_name]
            report_data["lines"]    = {int(k):v for k,v in report_data["lines"].items()}
            report_data["branches"] = {int(k):v for k,v in report_data["branches"].items()}

def parseAndCombine(paths):
    base_report = {}

    for path in paths:
        if path.endswith(".json"):
            with open(path) as f:
                report = json.load(f)
        elif path.endswith(".info"):
            report = parseInfo(path)
        else:
            logging.error("Currently only fastcov .json and lcov .info supported for combine operations, aborting due to %s...\n", path)
            sys.exit(EXIT_CODES["unsupported_coverage_format"])

        # In order for sorting to work later when we serialize,
        # make sure integer keys are int
        convertKeysToInt(report)

        if not base_report:
            base_report = report
            logging.info("Setting {} as base report".format(path))
        else:
            combineReports(base_report, report)
            logging.info("Adding {} to base report".format(path))

    return base_report

def getCombineCoverage(args):
    logging.info("Performing combine operation")
    fastcov_json = parseAndCombine(args.combine)
    filterFastcov(fastcov_json, args)
    return fastcov_json

def getGcovCoverage(args):
    # Need at least python 3.5 because of use of recursive glob
    checkPythonVersion(sys.version_info[0:2])

    # Need at least gcov 9.0.0 because that's when gcov JSON and stdout streaming was introduced
    checkGcovVersion(getGcovVersion(args.gcov))

    # Get list of gcda files to process
    coverage_files = findCoverageFiles(args.directory, args.coverage_files, args.use_gcno)

    # If gcda/gcno filtering is enabled, filter them out now
    if args.excludepre:
        coverage_files = getFilteredCoverageFiles(coverage_files, args.excludepre)
        logging.info("Found {} coverage files after filtering".format(len(coverage_files)))

    # We "zero" the "counters" by simply deleting all gcda files
    if args.zerocounters:
        removeFiles(globCoverageFiles(args.directory, GCOV_GCDA_EXT))
        logging.info("Removed {} .gcda files".format(len(coverage_files)))
        sys.exit()

    # Fire up one gcov per cpu and start processing gcdas
    gcov_filter_options = getGcovFilterOptions(args)
    fastcov_json = processGcdas(args, coverage_files, gcov_filter_options)

    # Summarize processing results
    logging.info("Processed {} .gcov files ({} total, {} skipped)".format(GCOVS_TOTAL - GCOVS_SKIPPED, GCOVS_TOTAL, GCOVS_SKIPPED))
    logging.debug("Final report will contain coverage for the following %d source files:\n    %s", len(fastcov_json["sources"]), "\n    ".join(fastcov_json["sources"]))

    return fastcov_json

def formatCoveredItems(covered, total):
    coverage = (covered * 100.0) / total if total > 0 else 100.0
    coverage = round(coverage, 2)
    return "{:.2f}%, {}/{}".format(coverage, covered, total)

def dumpStatistic(fastcov_json):
    total_lines = 0
    covered_lines = 0
    total_functions = 0
    covered_functions = 0
    total_files = len(fastcov_json["sources"])
    covered_files = 0

    for source_name, source in fastcov_json["sources"].items():
        is_file_covered = False
        for test_name, test in source.items():
            total_lines += len(test["lines"])
            for execution_count in test["lines"].values():
                covered_lines += 1 if execution_count > 0 else 0
                is_file_covered = is_file_covered or execution_count > 0
            total_functions += len(test["functions"])
            for function in test["functions"].values():
                covered_functions += 1 if function['execution_count'] > 0 else 0
                is_file_covered = is_file_covered or function['execution_count'] > 0
        if is_file_covered:
            covered_files = covered_files + 1

    logging.info("Files Coverage: {}".format(formatCoveredItems(covered_files, total_files)))
    logging.info("Functions Coverage: {}".format(formatCoveredItems(covered_functions, total_functions)))
    logging.info("Lines Coverage: {}".format(formatCoveredItems(covered_lines, total_lines)))

def dumpFile(fastcov_json, args):
    if args.lcov:
        dumpToLcovInfo(fastcov_json, args.output)
        logging.info("Created lcov info file '{}'".format(args.output))
    else:
        dumpToJson(fastcov_json, args.output)
        logging.info("Created fastcov json file '{}'".format(args.output))

    if args.dump_statistic:
        dumpStatistic(fastcov_json)

def tupleToDotted(tup):
    return ".".join(map(str, tup))

def parseArgs():
    parser = argparse.ArgumentParser(description='A parallel gcov wrapper for fast coverage report generation')
    parser.add_argument('-z', '--zerocounters', dest='zerocounters', action="store_true", help='Recursively delete all gcda files')

    # Enable Branch Coverage
    parser.add_argument('-b', '--branch-coverage', dest='branchcoverage', action="store_true", help='Include only the most useful branches in the coverage report.')
    parser.add_argument('-B', '--exceptional-branch-coverage', dest='xbranchcoverage', action="store_true", help='Include ALL branches in the coverage report (including potentially noisy exceptional branches).')
    parser.add_argument('-A', '--exclude-br-lines-starting-with', dest='exclude_branches_sw', nargs="+", metavar='', default=[], help='Exclude branches from lines starting with one of the provided strings (i.e. assert, return, etc.)')
    parser.add_argument('-a', '--include-br-lines-starting-with', dest='include_branches_sw', nargs="+", metavar='', default=[], help='Include only branches from lines starting with one of the provided strings (i.e. if, else, while, etc.)')
    parser.add_argument('-X', '--skip-exclusion-markers', dest='skip_exclusion_markers', action="store_true", help='Skip reading source files to search for lcov exclusion markers (such as "LCOV_EXCL_LINE")')
    parser.add_argument('-x', '--scan-exclusion-markers', dest='scan_exclusion_markers', action="store_true", help='(Combine operations) Force reading source files to search for lcov exclusion markers (such as "LCOV_EXCL_LINE")')

    # Capture untested file coverage as well via gcno
    parser.add_argument('-n', '--process-gcno', dest='use_gcno', action="store_true", help='Process both gcno and gcda coverage files. This option is useful for capturing untested files in the coverage report.')

    # Filtering Options
    parser.add_argument('-s', '--source-files', dest='sources',     nargs="+", metavar='', default=[], help='Filter: Specify exactly which source files should be included in the final report. Paths must be either absolute or relative to current directory.')
    parser.add_argument('-e', '--exclude',      dest='excludepost', nargs="+", metavar='', default=[], help='Filter: Exclude source files from final report if they contain one of the provided substrings (i.e. /usr/include test/, etc.)')
    parser.add_argument('-eg', '--exclude-glob', dest='excludepost_glob', nargs="+", metavar='', default=[], help='Filter: Exclude source files by glob pattern from final report if they contain one of the provided substrings (i.e. /usr/include test/, etc.)')
    parser.add_argument('-i', '--include',      dest='includepost', nargs="+", metavar='', default=[], help='Filter: Only include source files in final report that contain one of the provided substrings (i.e. src/ etc.)')
    parser.add_argument('-f', '--gcda-files',   dest='coverage_files',  nargs="+", metavar='', default=[], help='Filter: Specify exactly which gcda or gcno files should be processed. Note that specifying gcno causes both gcno and gcda to be processed.')
    parser.add_argument('-E', '--exclude-gcda', dest='excludepre',  nargs="+", metavar='', default=[], help='Filter: Exclude gcda or gcno files from being processed via simple find matching (not regex)')
    parser.add_argument('-u', '--diff-filter', dest='diff_file', default='', help='Unified diff file with changes which will be included into final report')
    parser.add_argument('-ub', '--diff-base-dir', dest='diff_base_dir', default='', help='Base directory for sources in unified diff file, usually repository dir')
    parser.add_argument('-ce', '--custom-exclusion-marker', dest='exclude_line_marker', nargs="+", metavar='', default=["LCOV_EXCL_LINE"], help='Filter: Add filter for lines that will be excluded from coverage (same behavior as "LCOV_EXCL_LINE")')

    parser.add_argument('-g', '--gcov', dest='gcov', default='gcov', help='Which gcov binary to use')

    parser.add_argument('-d', '--search-directory', dest='directory', default=".", help='Base directory to recursively search for gcda files (default: .)')
    parser.add_argument('-c', '--compiler-directory', dest='cdirectory', default="", help='Base directory compiler was invoked from (default: . or read from gcov) \
                                                                                            This needs to be set if invoking fastcov from somewhere other than the base compiler directory. No need to set it if gcc version > 9.1')

    parser.add_argument('-j', '--jobs', dest='jobs', type=int, default=multiprocessing.cpu_count(), help='Number of parallel gcov to spawn (default: {}).'.format(multiprocessing.cpu_count()))
    parser.add_argument('-m', '--minimum-chunk-size', dest='minimum_chunk', type=int, default=5,    help='Minimum number of files a thread should process (default: 5). \
                                                                                                          If you have only 4 gcda files but they are monstrously huge, you could change this value to a 1 so that each thread will only process 1 gcda. Otherwise fastcov will spawn only 1 thread to process all of them.')

    parser.add_argument('-F', '--fallback-encodings', dest='fallback_encodings', nargs="+", metavar='', default=[], help='List of encodings to try if opening a source file with the default fails (i.e. latin1, etc.). This option is not usually needed.')

    parser.add_argument('-l', '--lcov',   dest='lcov',   action="store_true",     help='Output in lcov info format instead of fastcov json')
    parser.add_argument('-o', '--output', dest='output', default="", help='Name of output file (default: coverage.json or coverage.info, depends on --lcov option)')
    parser.add_argument('-q', '--quiet',  dest='quiet',  action="store_true",     help='Suppress output to stdout')

    parser.add_argument('-t', '--test-name',     dest='test_name', default="", help='Specify a test name for the coverage. Equivalent to lcov\'s `-t`.')
    parser.add_argument('-C', '--add-tracefile', dest='combine',   nargs="+",  help='Combine multiple coverage files into one. If this flag is specified, fastcov will do a combine operation instead invoking gcov. Equivalent to lcov\'s `-a`.')

    parser.add_argument('-V', '--verbose', dest="verbose", action="store_true", help="Print more detailed information about what fastcov is doing")
    parser.add_argument('-w', '--validate-sources', dest="validate_sources", action="store_true", help="Check if every source file exists")
    parser.add_argument('-p', '--dump-statistic', dest="dump_statistic", action="store_true", help="Dump total statistic at the end")
    parser.add_argument('-v', '--version', action="version", version='%(prog)s {version}'.format(version=__version__), help="Show program's version number and exit")

    parser.add_argument('-gps', '--gcov_prefix_strip', dest="gcov_prefix_strip", action="store", default=0, type=int, help="The number of initial directory names to strip off the absolute paths in the object file.")
    parser.add_argument('-gp', '--gcov_prefix', dest="gcov_prefix", action="store", default="", help="The prefix to add to the paths in the object file.")

    args = parser.parse_args()
    if not args.output:
        args.output = 'coverage.info' if args.lcov else 'coverage.json'
    return args

def checkPythonVersion(version):
    """Exit if the provided python version is less than the supported version."""
    if version < MINIMUM_PYTHON:
        sys.stderr.write("Minimum python version {} required, found {}\n".format(tupleToDotted(MINIMUM_PYTHON), tupleToDotted(version)))
        sys.exit(EXIT_CODES["python_version"])

def checkGcovVersion(version):
    """Exit if the provided gcov version is less than the supported version."""
    if version < MINIMUM_GCOV:
        sys.stderr.write("Minimum gcov version {} required, found {}\n".format(tupleToDotted(MINIMUM_GCOV), tupleToDotted(version)))
        sys.exit(EXIT_CODES["gcov_version"])

def setupLogging(quiet, verbose):
    handler = logging.StreamHandler()
    handler.setFormatter(FastcovFormatter("[%(levelname)s]: %(message)s"))

    root = logging.getLogger()
    root.setLevel(logging.INFO)
    root.addHandler(handler)

    if not quiet:
        logging.disable(level=logging.NOTSET) # Re-enable logging

    if verbose:
        root.setLevel(logging.DEBUG)

def main():
    args = parseArgs()

    # Setup logging
    setupLogging(args.quiet, args.verbose)

    if args.gcov_prefix_strip > 0:
        os.environ["GCOV_PREFIX_STRIP"] = str(args.gcov_prefix_strip)

    if len(args.gcov_prefix) > 0:
        os.environ["GCOV_PREFIX"] = args.gcov_prefix

    # Get report from appropriate source
    if args.combine:
        fastcov_json = getCombineCoverage(args)
        skip_exclusion_markers = not args.scan_exclusion_markers
    else:
        fastcov_json = getGcovCoverage(args)
        skip_exclusion_markers = args.skip_exclusion_markers

    # Scan for exclusion markers
    if not skip_exclusion_markers:
        processExclusionMarkers(fastcov_json, args.jobs, args.exclude_branches_sw, args.include_branches_sw, args.exclude_line_marker, args.minimum_chunk, args.fallback_encodings, args.gcov_prefix, args.gcov_prefix_strip)
        logging.info("Scanned {} source files for exclusion markers".format(len(fastcov_json["sources"])))

    if args.diff_file:
        logging.info("Filtering according to {} file".format(args.diff_file))
        DiffParser().filterByDiff(args.diff_file, args.diff_base_dir, fastcov_json, args.fallback_encodings)

    if args.validate_sources:
        validateSources(fastcov_json, args.gcov_prefix, args.gcov_prefix_strip)

    # Dump to desired file format
    dumpFile(fastcov_json, args)

    # If there was an error along the way, but we still completed the pipeline...
    if EXIT_CODE:
        sys.exit(EXIT_CODE)

# Set package version... it's way down here so that we can call tupleToDotted
__version__ = tupleToDotted(FASTCOV_VERSION)

if __name__ == '__main__':
    main()