qpdf/check_abi

#!/usr/bin/env python3
import os
import sys
import argparse
import subprocess
import re

whoami = os.path.basename(sys.argv[0])
whereami = os.path.dirname(os.path.realpath(__file__))


def warn(*args, **kwargs):
    print(*args, file=sys.stderr, **kwargs)


class Main:
    def main(self, args=sys.argv[1:], prog=whoami):
        options = self.parse_args(args, prog)
        if options.action == 'dump':
            self.dump(options)
        elif options.action == 'check-sizes':
            self.check_sizes(options)
        elif options.action == 'compare':
            self.compare(options)
        else:
            exit(f'{whoami}: unknown action')

    def parse_args(self, args, prog):
        parser = argparse.ArgumentParser(
            prog=prog,
            # formatter_class=argparse.RawDescriptionHelpFormatter,
            description='Check ABI for changes',
        )

        subparsers = parser.add_subparsers(
            dest='action',
            help='specify subcommand; run action with --help for details',
            required=True)
        lib_arg = ('--lib', {'help': 'library file', 'required': True})

        p_dump = subparsers.add_parser(
            'dump',
            help='dump qpdf symbols in a library')
        p_dump.add_argument(lib_arg[0], **lib_arg[1])

        p_check_sizes = subparsers.add_parser(
            'check-sizes',
            help='check consistency between library and sizes.cc')
        p_check_sizes.add_argument(lib_arg[0], **lib_arg[1])

        p_compare = subparsers.add_parser(
            'compare',
            help='compare libraries and sizes')
        p_compare.add_argument('--new-lib',
                               help='new library file',
                               required=True)
        p_compare.add_argument('--old-lib',
                               help='old library file',
                               required=True)
        p_compare.add_argument('--old-sizes',
                               help='output of old sizes',
                               required=True)
        p_compare.add_argument('--new-sizes',
                               help='output of new sizes',
                               required=True)
        return parser.parse_args(args)

    def get_versions(self, path):
        p = os.path.basename(os.path.realpath(path))
        m = re.match(r'^libqpdf.so.(\d+).(\d+).(\d+)$', p)
        if not m:
            exit(f'{whoami}: {path} does end with libqpdf.so.x.y.z')
        major = int(m.group(1))
        minor = int(m.group(2))
        patch = int(m.group(3))
        return (major, minor, patch)

    def get_symbols(self, path):
        symbols = set()
        p = subprocess.run(
            ['nm', '-D', '--demangle', '--with-symbol-versions', path],
            stdout=subprocess.PIPE)
        if p.returncode:
            exit(f'{whoami}: failed to get symbols from {path}')
        for line in p.stdout.decode().split('\n'):
            # The LIBQPDF_\d+ comes from the version tag in
            # libqpdf.map.in.
            m = re.match(r'^[0-9a-f]+ (.) (.+)@@LIBQPDF_\d+\s*$', line)
            if not m:
                continue
            symbol = m.group(2)
            if re.match(r'^((void|int|bool|(.*? for)) )?std::', symbol):
                # Calling different methods of STL classes causes
                # different template instantiations to appear.
                # Standard library methods that sneak into the binary
                # interface are not considered part of the qpdf ABI.
                continue
            symbols.add(symbol)
        return symbols

    def dump(self, options):
        # This is just for manual use to investigate surprises.
        for i in sorted(self.get_symbols(options.lib)):
            print(i)

    def check_sizes(self, options):
        # Make sure that every class with methods in the public API
        # appears in sizes.cc either explicitly ignored or in a
        # print_size call. This enables us to reliably test whether
        # any object changed size by following the ABI checking
        # procedures outlined in README-maintainer.

        # To keep things up to date, whenever we add or remove
        # objects, we have to update sizes.cc. The check-sizes option
        # can be run at any time on an up-to-date build.

        lib = self.get_symbols(options.lib)
        classes = set()
        for i in sorted(lib):
            # Find a symbol that looks like a class method.
            m = re.match(
                r'(((?:^\S*?::)?(?:[^:\s]+))::([^:\s]+))(?:\[[^\]]+\])?\(',
                i)
            if m:
                full = m.group(1)
                clas = m.group(2)
                method = m.group(3)
                if full.startswith('std::') or method.startswith('~'):
                    # Sometimes std:: template instantiations make it
                    # into the library. Ignore those. Also ignore
                    # classes whose only exported method is a
                    # destructor.
                    continue
                # Otherwise, if the class exports a method, we
                # potentially care about changes to its size, so add
                # it.
                classes.add(clas)
        in_sizes = set()
        # Read the sizes.cc to make sure everything's there.
        with open(os.path.join(whereami, 'qpdf/sizes.cc'), 'r') as f:
            for line in f.readlines():
                m = re.search(r'^\s*(?:ignore_class|print_size)\((.*?)\)',
                              line)
                if m:
                    in_sizes.add(m.group(1))
        sizes_only = in_sizes - classes
        classes_only = classes - in_sizes
        if sizes_only or classes_only:
            if sizes_only:
                print("classes in sizes.cc but not in the library:")
                for i in sorted(sizes_only):
                    print(' ', i)
            if classes_only:
                print("classes in the library but not in sizes.cc:")
                for i in sorted(classes_only):
                    print(' ', i)
            exit(f'{whoami}: mismatch between library and sizes.cc')
        else:
            print(f'{whoami}: sizes.cc is consistent with the library')

    def read_sizes(self, filename):
        sizes = {}
        with open(filename, 'r') as f:
            for line in f.readlines():
                line = line.strip()
                m = re.match(r'^(.*) (\d+)$', line)
                if not m:
                    exit(f'{filename}: bad sizes line: {line}')
                sizes[m.group(1)] = m.group(2)
        return sizes

    def compare(self, options):
        old_version = self.get_versions(options.old_lib)
        new_version = self.get_versions(options.new_lib)
        old = self.get_symbols(options.old_lib)
        new = self.get_symbols(options.new_lib)
        if old_version > new_version:
            exit(f'{whoami}: old version is newer than new version')
        allow_abi_change = new_version[0] > old_version[0]
        allow_added = allow_abi_change or (new_version[1] > old_version[1])
        removed = sorted(old - new)
        added = sorted(new - old)
        if removed:
            print('INTERFACES REMOVED:')
            for i in removed:
                print(' ', i)
        else:
            print('No interfaces were removed')
        if added:
            print('INTERFACES ADDED')
            for i in added:
                print(' ', i)
        else:
            print('No interfaces were added')

        if removed and not allow_abi_change:
            exit(f'{whoami}: **ERROR**: major version must be bumped')
        elif added and not allow_added:
            exit(f'{whoami}: **ERROR**: minor version must be bumped')
        else:
            print(f'{whoami}: ABI check passed.')

        old_sizes = self.read_sizes(options.old_sizes)
        new_sizes = self.read_sizes(options.new_sizes)
        size_errors = False
        for k, v in old_sizes.items():
            if k in new_sizes and v != new_sizes[k]:
                size_errors = True
                print(f'{k} changed size from {v} to {new_sizes[k]}')
        if size_errors:
            if not allow_abi_change:
                exit(f'{whoami}:'
                     'size changes detected; this is an ABI change.')
        else:
            print(f'{whoami}: no size changes detected')


if __name__ == '__main__':
    try:
        Main().main()
    except KeyboardInterrupt:
        exit(130)