儿童网站开发/重庆seo网站排名
1、下载 OpenSSL 3.X 库的源代码放到待编译目录
2、解压并接入 OpenSSL 3.X 库源码的根目录
3、复制 ./Configure 一个取名为 ./Configure-clang
4、修改 ./Configure-clang
找到配置段:
CC=
CXX=
CPP=
LD=
把它们改成
CC => "/usr/bin/clang-8",
CXX => "/usr/bin/clang++-8",
CPP => "/usr/bin/clang++-8",
LD => "/usr/bin/clang++-8",
OpenSSL 3.X最好用 clang8 来编译,这是因为 OpenSSL 3.X 用了AVX512 指令集(汇编)本人试过 clang6 是无法编译OpenSSL使用的AVX512指令集汇编的。
完整配置参考:
#! /usr/bin/env perl
# -*- mode: perl; -*-
# Copyright 2016-2023 The OpenSSL Project Authors. All Rights Reserved.
#
# Licensed under the Apache License 2.0 (the "License"). You may not use
# this file except in compliance with the License. You can obtain a copy
# in the file LICENSE in the source distribution or at
# https://www.openssl.org/source/license.html## Configure -- OpenSSL source tree configuration scriptuse 5.10.0;
use strict;
use Config;
use FindBin;
use lib "$FindBin::Bin/util/perl";
use File::Basename;
use File::Spec::Functions qw/:DEFAULT abs2rel rel2abs splitdir/;
use File::Path qw/mkpath/;
use OpenSSL::fallback "$FindBin::Bin/external/perl/MODULES.txt";
use OpenSSL::Glob;
use OpenSSL::Template;
use OpenSSL::config;# see INSTALL.md for instructions.my $orig_death_handler = $SIG{__DIE__};
$SIG{__DIE__} = \&death_handler;my $usage="Usage: Configure [no-<feature> ...] [enable-<feature> ...] [-Dxxx] [-lxxx] [-Lxxx] [-fxxx] [-Kxxx] [no-hw-xxx|no-hw] [[no-]threads] [[no-]thread-pool] [[no-]default-thread-pool] [[no-]shared] [[no-]zlib|zlib-dynamic] [no-asm] [no-egd] [sctp] [386] [--prefix=DIR] [--openssldir=OPENSSLDIR] [--with-xxx[=vvv]] [--config=FILE] os/compiler[:flags]\n";my $banner = <<"EOF";**********************************************************************
*** ***
*** OpenSSL has been successfully configured ***
*** ***
*** If you encounter a problem while building, please open an ***
*** issue on GitHub <https://github.com/openssl/openssl/issues> ***
*** and include the output from the following command: ***
*** ***
*** perl configdata.pm --dump ***
*** ***
*** (If you are new to OpenSSL, you might want to consult the ***
*** 'Troubleshooting' section in the INSTALL.md file first) ***
*** ***
**********************************************************************
EOF# Options:
#
# --config add the given configuration file, which will be read after
# any "Configurations*" files that are found in the same
# directory as this script.
# --prefix prefix for the OpenSSL installation, which includes the
# directories bin, lib, include, share/man, share/doc/openssl
# This becomes the value of INSTALLTOP in Makefile
# (Default: /usr/local)
# --openssldir OpenSSL data area, such as openssl.cnf, certificates and keys.
# If it's a relative directory, it will be added on the directory
# given with --prefix.
# This becomes the value of OPENSSLDIR in Makefile and in C.
# (Default: PREFIX/ssl)
# --banner=".." Output specified text instead of default completion banner
#
# -w Don't wait after showing a Configure warning
#
# --cross-compile-prefix Add specified prefix to binutils components.
#
# --api One of 0.9.8, 1.0.0, 1.0.1, 1.0.2, 1.1.0, 1.1.1, or 3.0
# Define the public APIs as they were for that version
# including patch releases. If 'no-deprecated' is also
# given, do not compile support for interfaces deprecated
# up to and including the specified OpenSSL version.
#
# no-hw-xxx do not compile support for specific crypto hardware.
# Generic OpenSSL-style methods relating to this support
# are always compiled but return NULL if the hardware
# support isn't compiled.
# no-hw do not compile support for any crypto hardware.
# [no-]threads [don't] try to create a library that is suitable for
# multithreaded applications (default is "threads" if we
# know how to do it)
# [no-]thread-pool
# [don't] allow thread pool functionality
# [no-]default-thread-pool
# [don't] allow default thread pool functionality
# [no-]shared [don't] try to create shared libraries when supported.
# [no-]pic [don't] try to build position independent code when supported.
# If disabled, it also disables shared and dynamic-engine.
# no-asm do not use assembler
# no-egd do not compile support for the entropy-gathering daemon APIs
# [no-]zlib [don't] compile support for zlib compression.
# zlib-dynamic Like "zlib", but the zlib library is expected to be a shared
# library and will be loaded at run-time by the OpenSSL library.
# sctp include SCTP support
# no-quic disable QUIC support
# no-uplink Don't build support for UPLINK interface.
# enable-weak-ssl-ciphers
# Enable weak ciphers that are disabled by default.
# 386 generate 80386 code in assembly modules
# no-sse2 disables IA-32 SSE2 code in assembly modules, the above
# mentioned '386' option implies this one
# no-<cipher> build without specified algorithm (dsa, idea, rc5, ...)
# -<xxx> +<xxx> All options which are unknown to the 'Configure' script are
# /<xxx> passed through to the compiler. Unix-style options beginning
# with a '-' or '+' are recognized, as well as Windows-style
# options beginning with a '/'. If the option contains arguments
# separated by spaces, then the URL-style notation %20 can be
# used for the space character in order to avoid having to quote
# the option. For example, -opt%20arg gets expanded to -opt arg.
# In fact, any ASCII character can be encoded as %xx using its
# hexadecimal encoding.
# -static while -static is also a pass-through compiler option (and
# as such is limited to environments where it's actually
# meaningful), it triggers a number configuration options,
# namely no-pic, no-shared and no-threads. It is
# argued that the only reason to produce statically linked
# binaries (and in context it means executables linked with
# -static flag, and not just executables linked with static
# libcrypto.a) is to eliminate dependency on specific run-time,
# a.k.a. libc version. The mentioned config options are meant
# to achieve just that. Unfortunately on Linux it's impossible
# to eliminate the dependency completely for openssl executable
# because of getaddrinfo and gethostbyname calls, which can
# invoke dynamically loadable library facility anyway to meet
# the lookup requests. For this reason on Linux statically
# linked openssl executable has rather debugging value than
# production quality.
#
# BN_LLONG use the type 'long long' in crypto/bn/bn.h
# RC4_CHAR use 'char' instead of 'int' for RC4_INT in crypto/rc4/rc4.h
# Following are set automatically by this script
#
# MD5_ASM use some extra md5 assembler,
# SHA1_ASM use some extra sha1 assembler, must define L_ENDIAN for x86
# RMD160_ASM use some extra ripemd160 assembler,
# SHA256_ASM sha256_block is implemented in assembler
# SHA512_ASM sha512_block is implemented in assembler
# AES_ASM AES_[en|de]crypt is implemented in assembler# Minimum warning options... any contributions to OpenSSL should at least
# get past these. Note that we only use these with C compilers, not with
# C++ compilers.# -DPEDANTIC complements -pedantic and is meant to mask code that
# is not strictly standard-compliant and/or implementation-specific,
# e.g. inline assembly, disregards to alignment requirements, such
# that -pedantic would complain about. Incidentally -DPEDANTIC has
# to be used even in sanitized builds, because sanitizer too is
# supposed to and does take notice of non-standard behaviour. Then
# -pedantic with pre-C9x compiler would also complain about 'long
# long' not being supported. As 64-bit algorithms are common now,
# it grew impossible to resolve this without sizeable additional
# code, so we just tell compiler to be pedantic about everything
# but 'long long' type.my @gcc_devteam_warn = qw(-DPEDANTIC -pedantic -Wno-long-long -DUNUSEDRESULT_DEBUG-Wall-Wmissing-declarations-Wextra-Wno-unused-parameter-Wno-missing-field-initializers-Wswitch-Wsign-compare-Wshadow-Wformat-Wno-type-limits-Wno-tautological-constant-out-of-range-compare-Wundef-Werror-Wmissing-prototypes-Wstrict-prototypes
);# These are used in addition to $gcc_devteam_warn when the compiler is clang.
# TODO(openssl-team): fix problems and investigate if (at least) the
# following warnings can also be enabled:
# -Wcast-align
# -Wunreachable-code -- no, too ugly/compiler-specific
# -Wlanguage-extension-token -- no, we use asm()
# -Wunused-macros -- no, too tricky for BN and _XOPEN_SOURCE etc
# -Wextended-offsetof -- no, needed in CMS ASN1 code
my @clang_devteam_warn = qw(-Wno-unknown-warning-option-Wswitch-default-Wno-parentheses-equality-Wno-language-extension-token-Wno-extended-offsetof-Wno-missing-braces-Wconditional-uninitialized-Wincompatible-pointer-types-discards-qualifiers-Wmissing-variable-declarations
);my @cl_devteam_warn = qw(/WX
);my $strict_warnings = 0;# As for $BSDthreads. Idea is to maintain "collective" set of flags,
# which would cover all BSD flavors. -pthread applies to them all,
# but is treated differently. OpenBSD expands is as -D_POSIX_THREAD
# -lc_r, which is sufficient. FreeBSD 4.x expands it as -lc_r,
# which has to be accompanied by explicit -D_THREAD_SAFE and
# sometimes -D_REENTRANT. FreeBSD 5.x expands it as -lc_r, which
# seems to be sufficient?
our $BSDthreads="-pthread -D_THREAD_SAFE -D_REENTRANT";#
# API compatibility name to version number mapping.
#
my $apitable = {# This table expresses when API additions or changes can occur.# The numbering used changes from 3.0 and on because we updated# (solidified) our version numbering scheme at that point.# From 3.0 and on, we internalise the given version number in decimal# as MAJOR * 10000 + MINOR * 100 + 0"3.0.0" => 30000,"3.0" => 30000,# Note that before 3.0, we didn't have the same version number scheme.# Still, the numbering we use here covers what we need."1.1.1" => 10101,"1.1.0" => 10100,"1.0.2" => 10002,"1.0.1" => 10001,"1.0.0" => 10000,"0.9.8" => 908,
};# For OpenSSL::config::get_platform
my %guess_opts = ();my $dryrun = 0;our %table = ();
our %config = ();
our %withargs = ();
our $now_printing; # set to current entry's name in print_table_entry# (todo: right thing would be to encapsulate name# into %target [class] and make print_table_entry# a method)# Forward declarations ################################################ read_config(filename)
#
# Reads a configuration file and populates %table with the contents
# (which the configuration file places in %targets).
sub read_config;# resolve_config(target)
#
# Resolves all the late evaluations, inheritances and so on for the
# chosen target and any target it inherits from.
sub resolve_config;# Information collection ############################################## Unified build supports separate build dir
my $srcdir = catdir(absolutedir(dirname($0))); # catdir ensures local syntax
my $blddir = catdir(absolutedir(".")); # catdir ensures local syntax# File::Spec::Unix doesn't detect case insensitivity, so we make sure to
# check if the source and build directory are really the same, and make
# them so. This avoids all kinds of confusion later on.
# We must check @File::Spec::ISA rather than using File::Spec->isa() to
# know if File::Spec ended up loading File::Spec::Unix.
$srcdir = $blddirif (grep(/::Unix$/, @File::Spec::ISA)&& samedir($srcdir, $blddir));my $dofile = abs2rel(catfile($srcdir, "util/dofile.pl"));my $local_config_envname = 'OPENSSL_LOCAL_CONFIG_DIR';$config{sourcedir} = abs2rel($srcdir, $blddir);
$config{builddir} = abs2rel($blddir, $blddir);
# echo -n 'holy hand grenade of antioch' | openssl sha256
$config{FIPSKEY} ='f4556650ac31d35461610bac4ed81b1a181b2d8a43ea2854cbae22ca74560813';# Collect reconfiguration information if needed
my @argvcopy=@ARGV;if (grep /^reconf(igure)?$/, @argvcopy) {die "reconfiguring with other arguments present isn't supported"if scalar @argvcopy > 1;if (-f "./configdata.pm") {my $file = "./configdata.pm";unless (my $return = do $file) {die "couldn't parse $file: $@" if $@;die "couldn't do $file: $!" unless defined $return;die "couldn't run $file" unless $return;}@argvcopy = defined($configdata::config{perlargv}) ?@{$configdata::config{perlargv}} : ();die "Incorrect data to reconfigure, please do a normal configuration\n"if (grep(/^reconf/,@argvcopy));$config{perlenv} = $configdata::config{perlenv} // {};} else {die "Insufficient data to reconfigure, please do a normal configuration\n";}
}$config{perlargv} = [ @argvcopy ];# Historical: if known directories in crypto/ have been removed, it means
# that those sub-systems are disabled.
# (the other option would be to removed them from the SUBDIRS statement in
# crypto/build.info)
# We reverse the input list for cosmetic purely reasons, to compensate that
# 'unshift' adds at the front of the list (i.e. in reverse input order).
foreach ( reverse sort( 'aes', 'aria', 'bf', 'camellia', 'cast', 'des', 'dh','dsa', 'ec', 'hmac', 'idea', 'md2', 'md5', 'mdc2','rc2', 'rc4', 'rc5', 'ripemd', 'seed', 'sha','sm2', 'sm3', 'sm4') ) {unshift @argvcopy, "no-$_" if ! -d catdir($srcdir, 'crypto', $_);
}# Collect version numbers
my %version = ();collect_information(collect_from_file(catfile($srcdir,'VERSION.dat')),qr/\s*(\w+)\s*=\s*(.*?)\s*$/ =>sub {# Only define it if there is a value at allif ($2 ne '') {my $k = $1;my $v = $2;# Some values are quoted. Trim the quotes$v = $1 if $v =~ /^"(.*)"$/;$version{uc $k} = $v;}},"OTHERWISE" =>sub { die "Something wrong with this line:\n$_\nin $srcdir/VERSION.dat" },);$config{major} = $version{MAJOR} // 'unknown';
$config{minor} = $version{MINOR} // 'unknown';
$config{patch} = $version{PATCH} // 'unknown';
$config{prerelease} =defined $version{PRE_RELEASE_TAG} ? "-$version{PRE_RELEASE_TAG}" : '';
$config{build_metadata} =defined $version{BUILD_METADATA} ? "+$version{BUILD_METADATA}" : '';
$config{shlib_version} = $version{SHLIB_VERSION} // 'unknown';
$config{release_date} = $version{RELEASE_DATE} // 'xx XXX xxxx';$config{version} = "$config{major}.$config{minor}.$config{patch}";
$config{full_version} = "$config{version}$config{prerelease}$config{build_metadata}";die "erroneous version information in VERSION.dat: ","$config{version}, $config{shlib_version}\n"unless (defined $version{MAJOR}&& defined $version{MINOR}&& defined $version{PATCH}&& defined $version{SHLIB_VERSION});# Collect target configurationsmy $pattern = catfile(dirname($0), "Configurations", "*.conf");
foreach (sort glob($pattern)) {&read_config($_);
}if (defined env($local_config_envname)) {if ($^O eq 'VMS') {# VMS environment variables are logical names,# which can be used as is$pattern = $local_config_envname . ':' . '*.conf';} else {$pattern = catfile(env($local_config_envname), '*.conf');}foreach (sort glob($pattern)) {&read_config($_);}
}# Save away perl command information
$config{perl_cmd} = $^X;
$config{perl_version} = $Config{version};
$config{perl_archname} = $Config{archname};$config{prefix}="";
$config{openssldir}="";
$config{processor}="";
$config{libdir}="";
my $auto_threads=1; # enable threads automatically? true by default
my $default_ranlib;# Known TLS and DTLS protocols
my @tls = qw(ssl3 tls1 tls1_1 tls1_2 tls1_3);
my @dtls = qw(dtls1 dtls1_2);# Explicitly known options that are possible to disable. They can
# be regexps, and will be used like this: /^no-${option}$/
# For developers: keep it sorted alphabeticallymy @disablables = ("acvp-tests","afalgeng","apps","argon2","aria","asan","asm","async","autoalginit","autoerrinit","autoload-config","bf","blake2","brotli","brotli-dynamic","buildtest-c++","bulk","cached-fetch","camellia","capieng","winstore","cast","chacha","cmac","cmp","cms","comp","crypto-mdebug","ct","default-thread-pool","deprecated","des","devcryptoeng","dgram","dh","docs","dsa","dso","dtls","dynamic-engine","ec","ec2m","ec_nistp_64_gcc_128","ecdh","ecdsa","ecx","egd","engine","err","external-tests","filenames","fips","fips-securitychecks","fuzz-afl","fuzz-libfuzzer","gost","http","idea","ktls","legacy","loadereng","makedepend","md2","md4","mdc2","module","msan","multiblock","nextprotoneg","ocb","ocsp","padlockeng","pic","pinshared","poly1305","posix-io","psk","quic","rc2","rc4","rc5","rdrand","rfc3779","rmd160","scrypt","sctp","secure-memory","seed","shared","siphash","siv","sm2","sm2-precomp","sm3","sm4","sock","srp","srtp","sse2","ssl","ssl-trace","static-engine","stdio","tests","tfo","thread-pool","threads","tls","trace","ts","ubsan","ui-console","unit-test","uplink","weak-ssl-ciphers","whirlpool","zlib","zlib-dynamic","zstd","zstd-dynamic",);
foreach my $proto ((@tls, @dtls)){push(@disablables, $proto);push(@disablables, "$proto-method") unless $proto eq "tls1_3";}# Internal disablables, for aliasing purposes. They serve no special
# purpose here, but allow scripts to get to know them through configdata.pm,
# where these are merged with @disablables.
# The actual aliasing mechanism is done via %disable_cascades
my @disablables_int = qw(crmf);my %deprecated_disablables = ("ssl2" => undef,"buf-freelists" => undef,"crypto-mdebug-backtrace" => undef,"hw" => "hw", # causes cascade, but no macro"hw-padlock" => "padlockeng","ripemd" => "rmd160","ui" => "ui-console","heartbeats" => undef,);# All of the following are disabled by default:our %disabled = ( # "what" => "comment""fips" => "default","asan" => "default","brotli" => "default","brotli-dynamic" => "default","buildtest-c++" => "default","crypto-mdebug" => "default","crypto-mdebug-backtrace" => "default","devcryptoeng" => "default","ec_nistp_64_gcc_128" => "default","egd" => "default","external-tests" => "default","fuzz-afl" => "default","fuzz-libfuzzer" => "default","ktls" => "default","md2" => "default","msan" => "default","rc5" => "default","sctp" => "default","ssl3" => "default","ssl3-method" => "default","tfo" => "default","trace" => "default","ubsan" => "default","unit-test" => "default","weak-ssl-ciphers" => "default","zlib" => "default","zlib-dynamic" => "default","zstd" => "default","zstd-dynamic" => "default",);# Note: => pair form used for aesthetics, not to truly make a hash table
my @disable_cascades = (# "what" => [ "cascade", ... ]"bulk" => [ "shared", "dso","aria", "async", "autoload-config","blake2", "bf", "camellia", "cast", "chacha","cmac", "cms", "cmp", "comp", "ct","des", "dgram", "dh", "dsa","ec", "engine","filenames","idea", "ktls","md4", "multiblock", "nextprotoneg","ocsp", "ocb", "poly1305", "psk","rc2", "rc4", "rmd160","seed", "siphash", "siv","sm3", "sm4", "srp","srtp", "ssl3-method", "ssl-trace","tfo","ts", "ui-console", "whirlpool","fips-securitychecks" ],sub { $config{processor} eq "386" }=> [ "sse2" ],"ssl" => [ "ssl3" ],"ssl3-method" => [ "ssl3" ],"zlib" => [ "zlib-dynamic" ],"brotli" => [ "brotli-dynamic" ],"zstd" => [ "zstd-dynamic" ],"des" => [ "mdc2" ],"ec" => [ "ec2m", "ecdsa", "ecdh", "sm2", "gost", "ecx" ],"dgram" => [ "dtls", "quic", "sctp" ],"sock" => [ "dgram", "tfo" ],"dtls" => [ @dtls ],sub { 0 == scalar grep { !$disabled{$_} } @dtls }=> [ "dtls" ],"tls" => [ @tls ],sub { 0 == scalar grep { !$disabled{$_} } @tls }=> [ "tls" ],"tls1_3" => [ "quic" ],"crypto-mdebug" => [ "crypto-mdebug-backtrace" ],"module" => [ "dynamic-engine", "fips" ],# Without shared libraries, dynamic engines aren't possible.# This is due to them having to link with libcrypto and register features# using the ENGINE functionality, and since that relies on global tables,# those *have* to be exactly the same as the ones accessed from the app,# which cannot be guaranteed if shared libraries aren't present.# (note that even with shared libraries, both the app and dynamic engines# must be linked with the same library)"shared" => [ "dynamic-engine", "uplink" ],"dso" => [ "dynamic-engine", "module" ],# Other modules don't necessarily have to link with libcrypto, so shared# libraries do not have to be a condition to produce those.# Without position independent code, there can be no shared libraries# or modules."pic" => [ "shared", "module" ],"engine" => [ "dynamic-engine", grep(/eng$/, @disablables) ],"dynamic-engine" => [ "loadereng" ],"hw" => [ "padlockeng" ],# no-autoalginit is only useful when building non-shared"autoalginit" => [ "shared", "apps", "fips" ],"stdio" => [ "apps", "capieng", "egd" ],"apps" => [ "tests" ],"tests" => [ "external-tests" ],"comp" => [ "zlib", "brotli", "zstd" ],"sm3" => [ "sm2" ],sub { !$disabled{"unit-test"} } => [ "heartbeats" ],sub { !$disabled{"msan"} } => [ "asm" ],"cmac" => [ "siv" ],"legacy" => [ "md2" ],"cmp" => [ "crmf" ],"fips" => [ "fips-securitychecks", "acvp-tests" ],"threads" => [ "thread-pool" ],"thread-pool" => [ "default-thread-pool" ],"blake2" => [ "argon2" ],"deprecated-3.0" => [ "engine", "srp" ],"http" => [ "ocsp" ]);# Avoid protocol support holes. Also disable all versions below N, if version
# N is disabled while N+1 is enabled.
#
my @list = (reverse @tls);
while ((my $first, my $second) = (shift @list, shift @list)) {last unless @list;push @disable_cascades, ( sub { !$disabled{$first} && $disabled{$second} }=> [ @list ] );unshift @list, $second;
}
my @list = (reverse @dtls);
while ((my $first, my $second) = (shift @list, shift @list)) {last unless @list;push @disable_cascades, ( sub { !$disabled{$first} && $disabled{$second} }=> [ @list ] );unshift @list, $second;
}# Explicit "no-..." options will be collected in %disabled along with the defaults.
# To remove something from %disabled, use "enable-foo".
# For symmetry, "disable-foo" is a synonym for "no-foo".# For the "make variables" CPPINCLUDES and CPPDEFINES, we support lists with
# platform specific list separators. Users from those platforms should
# recognise those separators from how you set up the PATH to find executables.
# The default is the Unix like separator, :, but as an exception, we also
# support the space as separator.
my $list_separator_re ={ VMS => qr/(?<!\^),/,MSWin32 => qr/(?<!\\);/ } -> {$^O} // qr/(?<!\\)[:\s]/;
# All the "make variables" we support
# Some get pre-populated for the sake of backward compatibility
# (we supported those before the change to "make variable" support.
my %user = (AR => env('AR'),ARFLAGS => [],AS => undef,ASFLAGS => [],CC => "/usr/bin/clang-8",CFLAGS => [ env('CFLAGS') || () ],CXX => "/usr/bin/clang++-8",CXXFLAGS => [ env('CXXFLAGS') || () ],CPP => "/usr/bin/clang++-8",CPPFLAGS => [ env('CPPFLAGS') || () ], # -D, -I, -Wp,CPPDEFINES => [], # Alternative for -DCPPINCLUDES => [], # Alternative for -ICROSS_COMPILE => env('CROSS_COMPILE'),HASHBANGPERL=> env('HASHBANGPERL') || env('PERL'),LD => "/usr/bin/clang-8",LDFLAGS => [ env('LDFLAGS') || () ], # -L, -Wl,LDLIBS => [ env('LDLIBS') || () ], # -lMT => undef,MTFLAGS => [],PERL => env('PERL') || ($^O ne "VMS" ? $^X : "perl"),RANLIB => env('RANLIB'),RC => env('RC') || env('WINDRES'),RCFLAGS => [ env('RCFLAGS') || () ],RM => undef,);
# Info about what "make variables" may be prefixed with the cross compiler
# prefix. This should NEVER mention any such variable with a list for value.
my @user_crossable = qw ( AR AS CC CXX CPP LD MT RANLIB RC );
# The same but for flags given as Configure options. These are *additional*
# input, as opposed to the VAR=string option that override the corresponding
# config target attributes
my %useradd = (ASFLAGS => [],CPPDEFINES => [],CPPINCLUDES => [],CPPFLAGS => [],CFLAGS => [],CXXFLAGS => [],LDFLAGS => [],LDLIBS => [],RCFLAGS => [],);my %user_synonyms = (HASHBANGPERL=> 'PERL',RC => 'WINDRES',);# Some target attributes have been renamed, this is the translation table
my %target_attr_translate =(ar => 'AR',as => 'AS',cc => 'CC',cxx => 'CXX',cpp => 'CPP',hashbangperl => 'HASHBANGPERL',ld => 'LD',mt => 'MT',ranlib => 'RANLIB',rc => 'RC',rm => 'RM',);# Initialisers coming from 'config' scripts
$config{defines} = [ split(/$list_separator_re/, env('__CNF_CPPDEFINES')) ];
$config{includes} = [ split(/$list_separator_re/, env('__CNF_CPPINCLUDES')) ];
$config{cppflags} = [ env('__CNF_CPPFLAGS') || () ];
$config{cflags} = [ env('__CNF_CFLAGS') || () ];
$config{cxxflags} = [ env('__CNF_CXXFLAGS') || () ];
$config{lflags} = [ env('__CNF_LDFLAGS') || () ];
$config{ex_libs} = [ env('__CNF_LDLIBS') || () ];$config{openssl_api_defines}=[];
$config{openssl_sys_defines}=[];
$config{openssl_feature_defines}=[];
$config{options}="";
$config{build_type} = "release";
my $target="";my %cmdvars = (); # Stores FOO='blah' type arguments
my %unsupported_options = ();
my %deprecated_options = ();
# If you change this, update apps/version.c
my @known_seed_sources = qw(getrandom devrandom os egd none rdcpu librandom);
my @seed_sources = ();
while (@argvcopy){$_ = shift @argvcopy;# Support env variable assignments among the optionsif (m|^(\w+)=(.+)?$|){$cmdvars{$1} = $2;# Every time a variable is given as a configuration argument,# it acts as a reset if the variable.if (exists $user{$1}){$user{$1} = ref $user{$1} eq "ARRAY" ? [] : undef;}#if (exists $useradd{$1})# {# $useradd{$1} = [];# }next;}# VMS is a case insensitive environment, and depending on settings# out of our control, we may receive options uppercased. Let's# downcase at least the part before any equal sign.if ($^O eq "VMS"){s/^([^=]*)/lc($1)/e;}# some people just can't read the instructions, clang people have to...s/^-no-(?!integrated-as)/no-/;# rewrite some options in "enable-..." forms /^-?-?shared$/enable-shared/;s /^sctp$/enable-sctp/;s /^threads$/enable-threads/;s /^zlib$/enable-zlib/;s /^zlib-dynamic$/enable-zlib-dynamic/;s /^fips$/enable-fips/;if (/^(no|disable|enable)-(.+)$/){my $word = $2;if ($word !~ m|hw(?:-.+)| # special treatment for hw regexp opt&& !exists $deprecated_disablables{$word}&& !grep { $word eq $_ } @disablables){$unsupported_options{$_} = 1;next;}}if (/^no-(.+)$/ || /^disable-(.+)$/){foreach my $proto ((@tls, @dtls)){if ($1 eq "$proto-method"){$disabled{"$proto"} = "option($proto-method)";last;}}if ($1 eq "dtls"){foreach my $proto (@dtls){$disabled{$proto} = "option(dtls)";}$disabled{"dtls"} = "option(dtls)";}elsif ($1 eq "ssl"){# Last one of its kind$disabled{"ssl3"} = "option(ssl)";}elsif ($1 eq "tls"){# XXX: Tests will fail if all SSL/TLS# protocols are disabled.foreach my $proto (@tls){$disabled{$proto} = "option(tls)";}}elsif ($1 eq "static-engine"){delete $disabled{"dynamic-engine"};}elsif ($1 eq "dynamic-engine"){$disabled{"dynamic-engine"} = "option";}elsif (exists $deprecated_disablables{$1}){$deprecated_options{$_} = 1;if (defined $deprecated_disablables{$1}){$disabled{$deprecated_disablables{$1}} = "option";}}elsif ($1 =~ m|hw(?:-.+)|) # deprecate hw options in regexp form{$deprecated_options{$_} = 1;}else{$disabled{$1} = "option";}# No longer an automatic choice$auto_threads = 0 if ($1 eq "threads");}elsif (/^enable-(.+)$/){if ($1 eq "static-engine"){$disabled{"dynamic-engine"} = "option";}elsif ($1 eq "dynamic-engine"){delete $disabled{"dynamic-engine"};}elsif ($1 eq "zlib-dynamic"){delete $disabled{"zlib"};}elsif ($1 eq "brotli-dynamic"){delete $disabled{"brotli"};}elsif ($1 eq "zstd-dynamic"){delete $disabled{"zstd"};}my $algo = $1;delete $disabled{$algo};# No longer an automatic choice$auto_threads = 0 if ($1 eq "threads");}elsif (/^-d$/) # From older 'config'{$config{build_type} = "debug";}elsif (/^-v$/) # From older 'config'{$guess_opts{verbose} = 1;}elsif (/^-w$/){$guess_opts{nowait} = 1;}elsif (/^-t$/) # From older 'config'{$dryrun = 1;}elsif (/^--strict-warnings$/){# Pretend that our strict flags is a C flag, and replace it# with the proper flags later onpush @{$useradd{CFLAGS}}, '--ossl-strict-warnings';$strict_warnings=1;}elsif (/^--debug$/){$config{build_type} = "debug";}elsif (/^--release$/){$config{build_type} = "release";}elsif (/^386$/){ $config{processor}=386; }elsif (/^rsaref$/){# No RSAref support any more since it's not needed.# The check for the option is there so scripts aren't# broken}elsif (m|^[-+/]|){if (/^--prefix=(.*)$/){$config{prefix}=$1;}elsif (/^--api=(.*)$/){my $api = $1;die "Unknown API compatibility level $api"unless defined $apitable->{$api};$config{api}=$apitable->{$api};}elsif (/^--libdir=(.*)$/){$config{libdir}=$1;}elsif (/^--openssldir=(.*)$/){$config{openssldir}=$1;}elsif (/^--with-zlib-lib=(.*)$/){$withargs{zlib_lib}=$1;}elsif (/^--with-zlib-include=(.*)$/){$withargs{zlib_include}=$1;}elsif (/^--with-brotli-lib=(.*)$/){$withargs{brotli_lib}=$1;}elsif (/^--with-brotli-include=(.*)$/){$withargs{brotli_include}=$1;}elsif (/^--with-zstd-lib=(.*)$/){$withargs{zstd_lib}=$1;}elsif (/^--with-zstd-include=(.*)$/){$withargs{zstd_include}=$1;}elsif (/^--with-fuzzer-lib=(.*)$/){$withargs{fuzzer_lib}=$1;}elsif (/^--with-fuzzer-include=(.*)$/){$withargs{fuzzer_include}=$1;}elsif (/^--with-rand-seed=(.*)$/){foreach my $x (split(m|,|, $1)){die "Unknown --with-rand-seed choice $x\n"if ! grep { $x eq $_ } @known_seed_sources;push @seed_sources, $x;}}elsif (/^--fips-key=(.*)$/){$user{FIPSKEY}=lc($1);die "Non-hex character in FIPS key\n"if $user{FIPSKEY} =~ /[^a-f0-9]/;die "FIPS key must have even number of characters\n"if length $1 & 1;die "FIPS key too long (64 bytes max)\n"if length $1 > 64;}elsif (/^--banner=(.*)$/){$banner = $1 . "\n";}elsif (/^--cross-compile-prefix=(.*)$/){$user{CROSS_COMPILE}=$1;}elsif (/^--config=(.*)$/){read_config $1;}elsif (/^-l(.*)$/){push @{$useradd{LDLIBS}}, $_;}elsif (/^-framework$/){push @{$useradd{LDLIBS}}, $_, shift(@argvcopy);}elsif (/^-L(.*)$/ or /^-Wl,/){push @{$useradd{LDFLAGS}}, $_;}elsif (/^-rpath$/ or /^-R$/)# -rpath is the OSF1 rpath flag# -R is the old Solaris rpath flag{my $rpath = shift(@argvcopy) || "";$rpath .= " " if $rpath ne "";push @{$useradd{LDFLAGS}}, $_, $rpath;}elsif (/^-static$/){push @{$useradd{LDFLAGS}}, $_;}elsif (m|^[-/]D(.*)$|){push @{$useradd{CPPDEFINES}}, $1;}elsif (m|^[-/]I(.*)$|){push @{$useradd{CPPINCLUDES}}, $1;}elsif (/^-Wp,$/){push @{$useradd{CPPFLAGS}}, $1;}else # common if (/^[-+]/), just pass down...{# Treat %xx as an ASCII code (e.g. replace %20 by a space character).# This provides a simple way to pass options with arguments separated# by spaces without quoting (e.g. -opt%20arg translates to -opt arg).$_ =~ s/%([0-9a-f]{1,2})/chr(hex($1))/gei;push @{$useradd{CFLAGS}}, $_;push @{$useradd{CXXFLAGS}}, $_;}}elsif (m|^/|){# Treat %xx as an ASCII code (e.g. replace %20 by a space character).# This provides a simple way to pass options with arguments separated# by spaces without quoting (e.g. /opt%20arg translates to /opt arg).$_ =~ s/%([0-9a-f]{1,2})/chr(hex($1))/gei;push @{$useradd{CFLAGS}}, $_;push @{$useradd{CXXFLAGS}}, $_;}else{die "target already defined - $target (offending arg: $_)\n" if ($target ne "");$target=$_;}unless ($_ eq $target || /^no-/ || /^disable-/){# "no-..." follows later after implied deactivations# have been derived. (Don't take this too seriously,# we really only write OPTIONS to the Makefile out of# nostalgia.)if ($config{options} eq ""){ $config{options} = $_; }else{ $config{options} .= " ".$_; }}}if (keys %deprecated_options){warn "***** Deprecated options: ",join(", ", keys %deprecated_options), "\n";}
if (keys %unsupported_options){die "***** Unsupported options: ",join(", ", keys %unsupported_options), "\n";}# If any %useradd entry has been set, we must check that the "make
# variables" haven't been set. We start by checking of any %useradd entry
# is set.
if (grep { scalar @$_ > 0 } values %useradd) {# Hash of env / make variables names. The possible values are:# 1 - "make vars"# 2 - %useradd entry set# 3 - both setmy %detected_vars =map { my $v = 0;$v += 1 if $cmdvars{$_};$v += 2 if @{$useradd{$_}};$_ => $v }keys %useradd;# If any of the corresponding "make variables" is set, we errorif (grep { $_ & 1 } values %detected_vars) {my $names = join(', ', grep { $detected_vars{$_} > 0 }sort keys %detected_vars);die <<"_____";
***** Mixing make variables and additional compiler/linker flags as
***** configure command line option is not permitted.
***** Affected make variables: $names
_____}
}# Check through all supported command line variables to see if any of them
# were set, and canonicalise the values we got. If no compiler or linker
# flag or anything else that affects %useradd was set, we also check the
# environment for values.
my $anyuseradd =grep { defined $_ && (ref $_ ne 'ARRAY' || @$_) } values %useradd;
foreach (keys %user) {my $value = $cmdvars{$_};$value //= env($_) unless $anyuseradd;$value //=defined $user_synonyms{$_} ? $cmdvars{$user_synonyms{$_}} : undef;$value //= defined $user_synonyms{$_} ? env($user_synonyms{$_}) : undefunless $anyuseradd;if (defined $value) {if (ref $user{$_} eq 'ARRAY') {if ($_ eq 'CPPDEFINES' || $_ eq 'CPPINCLUDES') {$user{$_} = [ split /$list_separator_re/, $value ];} else {$user{$_} = [ $value ];}} elsif (!defined $user{$_}) {$user{$_} = $value;}}
}if (grep { /-rpath\b/ } ($user{LDFLAGS} ? @{$user{LDFLAGS}} : ())&& !$disabled{shared}&& !($disabled{asan} && $disabled{msan} && $disabled{ubsan})) {die "***** Cannot simultaneously use -rpath, shared libraries, and\n","***** any of asan, msan or ubsan\n";
}# If no target was given, try guessing.
unless ($target) {my %system_config = OpenSSL::config::get_platform(%guess_opts, %user);# The $system_config{disable} is used to populate %disabled with# entries that aren't already there.foreach ( @{$system_config{disable} // []} ) {$disabled{$_} = 'system' unless defined $disabled{$_};}delete $system_config{disable};# Override config entries with stuff from the guesser.# It's assumed that this really is nothing new.%config = ( %config, %system_config );$target = $system_config{target};
}sub disable {my $disable_type = shift;for (@_) {$disabled{$_} = $disable_type;}my @tocheckfor = (@_ ? @_ : keys %disabled);while (@tocheckfor) {my %new_tocheckfor = ();my @cascade_copy = (@disable_cascades);while (@cascade_copy) {my ($test, $descendents) =(shift @cascade_copy, shift @cascade_copy);if (ref($test) eq "CODE" ? $test->() : defined($disabled{$test})) {foreach (grep { !defined($disabled{$_}) } @$descendents) {$new_tocheckfor{$_} = 1; $disabled{$_} = "cascade";}}}@tocheckfor = (keys %new_tocheckfor);}
}
disable(); # First cascade runour $die = sub { die @_; };
if ($target eq "TABLE") {local $die = sub { warn @_; };foreach (sort keys %table) {print_table_entry($_, "TABLE");}exit 0;
}if ($target eq "LIST") {foreach (sort keys %table) {print $_,"\n" unless $table{$_}->{template};}exit 0;
}if ($target eq "HASH") {local $die = sub { warn @_; };print "%table = (\n";foreach (sort keys %table) {print_table_entry($_, "HASH");}exit 0;
}print "Configuring OpenSSL version $config{full_version} ";
print "for target $target\n";if (scalar(@seed_sources) == 0) {print "Using os-specific seed configuration\n";push @seed_sources, 'os';
}
if (scalar(grep { $_ eq 'egd' } @seed_sources) > 0) {delete $disabled{'egd'};
}
if (scalar(grep { $_ eq 'none' } @seed_sources) > 0) {die "Cannot seed with none and anything else" if scalar(@seed_sources) > 1;warn <<_____ if scalar(@seed_sources) == 1;============================== WARNING ===============================
You have selected the --with-rand-seed=none option, which effectively
disables automatic reseeding of the OpenSSL random generator.
All operations depending on the random generator such as creating keys
will not work unless the random generator is seeded manually by the
application.Please read the 'Note on random number generation' section in the
INSTALL.md instructions and the RAND_DRBG(7) manual page for more
details.
============================== WARNING ===============================_____
}
push @{$config{openssl_feature_defines}},map { (my $x = $_) =~ tr|[\-a-z]|[_A-Z]|; "OPENSSL_RAND_SEED_$x" }@seed_sources;# Backward compatibility?
if ($target =~ m/^CygWin32(-.*)$/) {$target = "Cygwin".$1;
}# Support for legacy targets having a name starting with 'debug-'
my ($d, $t) = $target =~ m/^(debug-)?(.*)$/;
if ($d) {$config{build_type} = "debug";# If we do not find debug-foo in the table, the target is set to foo.if (!$table{$target}) {$target = $t;}
}if ($target) {# It's possible that we have different config targets for specific# toolchains, so we try to detect them, and go for the plain config# target if not.my $found;foreach ( ( "$target-$user{CC}", "$target", undef ) ) {$found=$_ if $table{$_} && !$table{$_}->{template};last if $found;}$target = $found;
} else {# If we don't have a config target now, we try the C compiler as we# fallbackmy $cc = $user{CC} // 'cc';$target = $cc if $table{$cc} && !$table{$cc}->{template};
}&usage unless $target;exit 0 if $dryrun; # From older 'config'$config{target} = $target;
my %target = resolve_config($target);foreach (keys %target_attr_translate) {$target{$target_attr_translate{$_}} = $target{$_}if $target{$_};delete $target{$_};
}%target = ( %{$table{DEFAULTS}}, %target );my %conf_files = map { $_ => 1 } (@{$target{_conf_fname_int}});
$config{conf_files} = [ sort keys %conf_files ];# Using sub disable within these loops may prove fragile, so we run
# a cascade afterwards
foreach my $feature (@{$target{disable}}) {if (exists $deprecated_disablables{$feature}) {warn "***** config $target disables deprecated feature $feature\n";} elsif (!grep { $feature eq $_ } @disablables) {die "***** config $target disables unknown feature $feature\n";}$disabled{$feature} = 'config';
}
foreach my $feature (@{$target{enable}}) {if ("default" eq ($disabled{$feature} // "")) {if (exists $deprecated_disablables{$feature}) {warn "***** config $target enables deprecated feature $feature\n";} elsif (!grep { $feature eq $_ } @disablables) {die "***** config $target enables unknown feature $feature\n";}delete $disabled{$feature};}
}# If uplink_arch isn't defined, disable uplink
$disabled{uplink} = 'no uplink_arch' unless (defined $target{uplink_arch});
# If asm_arch isn't defined, disable asm
$disabled{asm} = 'no asm_arch' unless (defined $target{asm_arch});disable(); # Run a cascade now$target{CXXFLAGS}//=$target{CFLAGS} if $target{CXX};
$target{cxxflags}//=$target{cflags} if $target{CXX};
$target{exe_extension}=".exe" if ($config{target} eq "DJGPP");
$target{exe_extension}=".pm" if ($config{target} =~ /vos/);# Fill %config with values from %user, and in case those are undefined or
# empty, use values from %target (acting as a default).
foreach (keys %user) {my $ref_type = ref $user{$_};# Temporary function. Takes an intended ref type (empty string or "ARRAY")# and a value that's to be coerced into that type.my $mkvalue = sub {my $type = shift;my $value = shift;my $undef_p = shift;die "Too many arguments for \$mkvalue" if @_;while (ref $value eq 'CODE') {$value = $value->();}if ($type eq 'ARRAY') {return undef unless defined $value;return undef if ref $value ne 'ARRAY' && !$value;return undef if ref $value eq 'ARRAY' && !@$value;return [ $value ] unless ref $value eq 'ARRAY';}return undef unless $value;return $value;};$config{$_} =$mkvalue->($ref_type, $user{$_})|| $mkvalue->($ref_type, $target{$_});delete $config{$_} unless defined $config{$_};
}# Finish up %config by appending things the user gave us on the command line
# apart from "make variables"
foreach (keys %useradd) {# The must all be lists, so we assert that heredie "internal error: \$useradd{$_} isn't an ARRAY\n"unless ref $useradd{$_} eq 'ARRAY';if (defined $config{$_}) {push @{$config{$_}}, @{$useradd{$_}};} else {$config{$_} = [ @{$useradd{$_}} ];}
}
# At this point, we can forget everything about %user and %useradd,
# because it's now all been merged into the corresponding $config entryif ($config{prefix} && !$config{CROSS_COMPILE}) {die "Directory given with --prefix MUST be absolute\n"unless file_name_is_absolute($config{prefix});
}if (grep { $_ =~ /(?:^|\s)-static(?:\s|$)/ } @{$config{LDFLAGS}}) {disable('static', 'pic', 'threads');
}# Allow overriding the build file name
$config{build_file} = env('BUILDFILE') || $target{build_file} || "Makefile";# Make sure build_scheme is consistent.
$target{build_scheme} = [ $target{build_scheme} ]if ref($target{build_scheme}) ne "ARRAY";my ($builder, $builder_platform, @builder_opts) =@{$target{build_scheme}};foreach my $checker (($builder_platform."-".$config{build_file}."-checker.pm",$builder_platform."-checker.pm")) {my $checker_path = catfile($srcdir, "Configurations", $checker);if (-f $checker_path) {my $fn = $ENV{CONFIGURE_CHECKER_WARN}? sub { warn $@; } : sub { die $@; };if (! do $checker_path) {if ($@) {$fn->($@);} elsif ($!) {$fn->($!);} else {$fn->("The detected tools didn't match the platform\n");}}last;}
}push @{$config{defines}}, "NDEBUG" if $config{build_type} eq "release";if ($target =~ /^mingw/ && `$config{CC} --target-help 2>&1` =~ m/-mno-cygwin/m){push @{$config{cflags}}, "-mno-cygwin";push @{$config{cxxflags}}, "-mno-cygwin" if $config{CXX};push @{$config{shared_ldflag}}, "-mno-cygwin";}if ($target =~ /linux.*-mips/ && !$disabled{asm}&& !grep { $_ =~ /-m(ips|arch=)/ } (@{$config{CFLAGS}})) {# minimally required architecture flags for assembly modulesmy $value;$value = '-mips2' if ($target =~ /mips32/);$value = '-mips3' if ($target =~ /mips64/);unshift @{$config{cflags}}, $value;unshift @{$config{cxxflags}}, $value if $config{CXX};
}# If threads aren't disabled, check how possible they are
unless ($disabled{threads}) {if ($auto_threads) {# Enabled by default, disable it forcibly if unavailableif ($target{thread_scheme} eq "(unknown)") {disable("unavailable", 'threads');}} else {# The user chose to enable threads explicitly, let's see# if there's a chance that's possibleif ($target{thread_scheme} eq "(unknown)") {# If the user asked for "threads" and we don't have internal# knowledge how to do it, [s]he is expected to provide any# system-dependent compiler options that are necessary. We# can't truly check that the given options are correct, but# we expect the user to know what [s]He is doing.if (!@{$config{CFLAGS}} && !@{$config{CPPDEFINES}}) {die "You asked for multi-threading support, but didn't\n","provide any system-specific compiler options\n";}}}
}# Find out if clang's sanitizers have been enabled with -fsanitize
# flags and ensure that the corresponding %disabled elements area
# removed to reflect that the sanitizers are indeed enabled.
my %detected_sanitizers = ();
foreach (grep /^-fsanitize=/, @{$config{CFLAGS} || []}) {(my $checks = $_) =~ s/^-fsanitize=//;foreach (split /,/, $checks) {my $d = { address => 'asan',undefined => 'ubsan',memory => 'msan' } -> {$_};next unless defined $d;$detected_sanitizers{$d} = 1;if (defined $disabled{$d}) {die "***** Conflict between disabling $d and enabling $_ sanitizer"if $disabled{$d} ne "default";delete $disabled{$d};}}
}# If threads still aren't disabled, add a C macro to ensure the source
# code knows about it. Any other flag is taken care of by the configs.
unless($disabled{threads}) {push @{$config{openssl_feature_defines}}, "OPENSSL_THREADS";
}my $no_shared_warn=0;
if (($target{shared_target} // '') eq ""){$no_shared_warn = 1if (!$disabled{shared} || !$disabled{"dynamic-engine"});disable('no-shared-target', 'pic');}if ($disabled{"dynamic-engine"}) {$config{dynamic_engines} = 0;
} else {$config{dynamic_engines} = 1;
}unless ($disabled{asan} || defined $detected_sanitizers{asan}) {push @{$config{cflags}}, "-fsanitize=address";
}unless ($disabled{ubsan} || defined $detected_sanitizers{ubsan}) {push @{$config{cflags}}, "-fsanitize=undefined", "-fno-sanitize-recover=all", "-DPEDANTIC";
}unless ($disabled{msan} || defined $detected_sanitizers{msan}) {push @{$config{cflags}}, "-fsanitize=memory";
}unless ($disabled{"fuzz-libfuzzer"} && $disabled{"fuzz-afl"}&& $disabled{asan} && $disabled{ubsan} && $disabled{msan}) {push @{$config{cflags}}, "-fno-omit-frame-pointer", "-g";push @{$config{cxxflags}}, "-fno-omit-frame-pointer", "-g" if $config{CXX};
}
#
# Platform fix-ups
## This saves the build files from having to check
if ($disabled{pic}){foreach (qw(shared_cflag shared_cxxflag shared_cppflagshared_defines shared_includes shared_ldflagmodule_cflags module_cxxflags module_cppflagsmodule_defines module_includes module_lflags)){delete $config{$_};$target{$_} = "";}}
else{push @{$config{lib_defines}}, "OPENSSL_PIC";}if ($target{sys_id} ne ""){push @{$config{openssl_sys_defines}}, "OPENSSL_SYS_$target{sys_id}";}my %predefined_C = compiler_predefined($config{CROSS_COMPILE}.$config{CC});
my %predefined_CXX = $config{CXX}? compiler_predefined($config{CROSS_COMPILE}.$config{CXX}): ();unless ($disabled{asm}) {# big endian systems can use ELFv2 ABIif ($target eq "linux-ppc64" || $target eq "BSD-ppc64") {$target{perlasm_scheme} = "linux64v2" if ($predefined_C{_CALL_ELF} == 2);}
}# Check for makedepend capabilities.
if (!$disabled{makedepend}) {# If the attribute makedep_scheme is defined, then we assume that the# config target and its associated build file are programmed to deal# with it.# If makedep_scheme is undefined, we go looking for GCC compatible# dependency making, and if that's not available, we try to fall back# on 'makedepend'.if ($target{makedep_scheme}) {$config{makedep_scheme} = $target{makedep_scheme};# If the makedepcmd attribute is defined, copy it. If not, the# build files will have to fend for themselves.$config{makedepcmd} = $target{makedepcmd} if $target{makedepcmd};} elsif (($predefined_C{__GNUC__} // -1) >= 3&& !($predefined_C{__APPLE_CC__} && !$predefined_C{__clang__})) {# We know that GNU C version 3 and up as well as all clang# versions support dependency generation, but Xcode did not# handle $cc -M before clang support (but claims __GNUC__ = 3)$config{makedep_scheme} = 'gcc';} else {# In all other cases, we look for 'makedepend', and set the# makedep_scheme value if we found it.$config{makedepcmd} = which('makedepend');$config{makedep_scheme} = 'makedepend' if $config{makedepcmd};}# If no depend scheme is set, we disable makedependdisable('unavailable', 'makedepend') unless $config{makedep_scheme};
}if (!$disabled{asm} && !$predefined_C{__MACH__} && $^O ne 'VMS') {# probe for -Wa,--noexecstack option...if ($predefined_C{__clang__}) {# clang has builtin assembler, which doesn't recognize --help,# but it apparently recognizes the option in question on all# supported platforms even when it's meaningless. In other words# probe would fail, but probed option always accepted...push @{$config{cflags}}, "-Wa,--noexecstack", "-Qunused-arguments";} else {my $cc = $config{CROSS_COMPILE}.$config{CC};open(PIPE, "$cc -Wa,--help -c -o null.$$.o -x assembler /dev/null 2>&1 |");while(<PIPE>) {if (m/--noexecstack/) {push @{$config{cflags}}, "-Wa,--noexecstack";last;}}close(PIPE);unlink("null.$$.o");}
}# Deal with bn_ops ###################################################$config{bn_ll} =0;
my $def_int="unsigned int";
$config{rc4_int} =$def_int;
($config{b64l},$config{b64},$config{b32})=(0,0,1);my $count = 0;
foreach (sort split(/\s+/,$target{bn_ops})) {$count++ if /SIXTY_FOUR_BIT|SIXTY_FOUR_BIT_LONG|THIRTY_TWO_BIT/;$config{bn_ll}=1 if $_ eq 'BN_LLONG';$config{rc4_int}="unsigned char" if $_ eq 'RC4_CHAR';($config{b64l},$config{b64},$config{b32})=(0,1,0) if $_ eq 'SIXTY_FOUR_BIT';($config{b64l},$config{b64},$config{b32})=(1,0,0) if $_ eq 'SIXTY_FOUR_BIT_LONG';($config{b64l},$config{b64},$config{b32})=(0,0,1) if $_ eq 'THIRTY_TWO_BIT';
}
die "Exactly one of SIXTY_FOUR_BIT|SIXTY_FOUR_BIT_LONG|THIRTY_TWO_BIT can be set in bn_ops\n"if $count > 1;$config{api} = $config{major} * 10000 + $config{minor} * 100unless $config{api};
foreach (keys %$apitable) {$disabled{"deprecated-$_"} = "deprecation"if $disabled{deprecated} && $config{api} >= $apitable->{$_};
}disable(); # Run a cascade now# Hack cflags for better warnings (dev option) ######################## "Stringify" the C and C++ flags string. This permits it to be made part of
# a string and works as well on command lines.
$config{cflags} = [ map { (my $x = $_) =~ s/([\\\"])/\\$1/g; $x }@{$config{cflags}} ];
$config{cxxflags} = [ map { (my $x = $_) =~ s/([\\\"])/\\$1/g; $x }@{$config{cxxflags}} ] if $config{CXX};$config{openssl_api_defines} = ["OPENSSL_CONFIGURED_API=".$config{api},
];my @strict_warnings_collection=();
if ($strict_warnings){my $wopt;my $gccver = $predefined_C{__GNUC__} // -1;if ($gccver >= 4){push @strict_warnings_collection, @gcc_devteam_warn;push @strict_warnings_collection, @clang_devteam_warnif (defined($predefined_C{__clang__}));}elsif ($config{target} =~ /^VC-/){push @strict_warnings_collection, @cl_devteam_warn;}else{warn "WARNING --strict-warnings requires gcc[>=4] or gcc-alike, or MSVC"}}$config{CFLAGS} = [ map { $_ eq '--ossl-strict-warnings'? @strict_warnings_collection: ( $_ ) }@{$config{CFLAGS}} ];unless ($disabled{afalgeng}) {$config{afalgeng}="";if (grep { $_ eq 'afalgeng' } @{$target{enable}}) {push @{$config{engdirs}}, "afalg";} else {disable('not-linux', 'afalgeng');}
}unless ($disabled{devcryptoeng}) {if ($target =~ m/^BSD/) {my $maxver = 5*100 + 7;my $sysstr = `uname -s`;my $verstr = `uname -r`;$sysstr =~ s|\R$||;$verstr =~ s|\R$||;my ($ma, $mi, @rest) = split m|\.|, $verstr;my $ver = $ma*100 + $mi;if ($sysstr eq 'OpenBSD' && $ver >= $maxver) {disable('too-new-kernel', 'devcryptoeng');}}
}unless ($disabled{ktls}) {$config{ktls}="";my $cc = $config{CROSS_COMPILE}.$config{CC};if ($target =~ m/^linux/) {system("printf '#include <sys/types.h>\n#include <linux/tls.h>' | $cc -E - >/dev/null 2>&1");if ($? != 0) {disable('too-old-kernel', 'ktls');}} elsif ($target =~ m/^BSD/) {system("printf '#include <sys/types.h>\n#include <sys/ktls.h>' | $cc -E - >/dev/null 2>&1");if ($? != 0) {disable('too-old-freebsd', 'ktls');}} else {disable('not-linux-or-freebsd', 'ktls');}
}unless ($disabled{winstore}) {unless ($target =~ /^(?:Cygwin|mingw|VC-|BC-)/) {disable('not-windows', 'winstore');}
}push @{$config{openssl_other_defines}}, "OPENSSL_NO_KTLS" if ($disabled{ktls});# Get the extra flags used when building shared libraries and modules. We
# do this late because some of them depend on %disabled.# Make the flags to build DSOs the same as for shared libraries unless they
# are already defined
$target{module_cflags} = $target{shared_cflag} unless defined $target{module_cflags};
$target{module_cxxflags} = $target{shared_cxxflag} unless defined $target{module_cxxflags};
$target{module_ldflags} = $target{shared_ldflag} unless defined $target{module_ldflags};
{my $shared_info_pl =catfile(dirname($0), "Configurations", "shared-info.pl");my %shared_info = read_eval_file($shared_info_pl);push @{$target{_conf_fname_int}}, $shared_info_pl;my $si = $target{shared_target};while (ref $si ne "HASH") {last if ! defined $si;if (ref $si eq "CODE") {$si = $si->();} else {$si = $shared_info{$si};}}# Some of the 'shared_target' values don't have any entries in# %shared_info. That's perfectly fine, AS LONG AS the build file# template knows how to handle this. That is currently the case for# Windows and VMS.if (defined $si) {# Just as above, copy certain shared_* attributes to the corresponding# module_ attribute unless the latter is already defined$si->{module_cflags} = $si->{shared_cflag} unless defined $si->{module_cflags};$si->{module_cxxflags} = $si->{shared_cxxflag} unless defined $si->{module_cxxflags};$si->{module_ldflags} = $si->{shared_ldflag} unless defined $si->{module_ldflags};foreach (sort keys %$si) {$target{$_} = defined $target{$_}? add($si->{$_})->($target{$_}): $si->{$_};}}
}# ALL MODIFICATIONS TO %disabled, %config and %target MUST BE DONE FROM HERE ON######################################################################
# Build up information for skipping certain directories depending on disabled
# features, as well as setting up macros for disabled features.# This is a tentative database of directories to skip. Some entries may not
# correspond to anything real, but that's ok, they will simply be ignored.
# The actual processing of these entries is done in the build.info lookup
# loop further down.
#
# The key is a Unix formatted path in the source tree, the value is an index
# into %disabled_info, so any existing path gets added to a corresponding
# 'skipped' entry in there with the list of skipped directories.
my %skipdir = ();
my %disabled_info = (); # For configdata.pm
foreach my $what (sort keys %disabled) {# There are deprecated disablables that translate to themselves.# They cause disabling cascades, but should otherwise not register.next if $deprecated_disablables{$what};# The generated $disabled{"deprecated-x.y"} entries are special# and treated properly elsewherenext if $what =~ m|^deprecated-|;$config{options} .= " no-$what";if (!grep { $what eq $_ } ( 'buildtest-c++', 'fips', 'threads', 'shared','module', 'pic', 'dynamic-engine', 'makedepend','sse2', 'legacy' )) {(my $WHAT = uc $what) =~ s|-|_|g;my $skipdir = $what;# fix-up crypto/directory name(s)$skipdir = "ripemd" if $what eq "rmd160";$skipdir = "whrlpool" if $what eq "whirlpool";my $macro = $disabled_info{$what}->{macro} = "OPENSSL_NO_$WHAT";push @{$config{openssl_feature_defines}}, $macro;$skipdir{engines} = $what if $what eq 'engine';$skipdir{"crypto/$skipdir"} = $whatunless $what eq 'async' || $what eq 'err' || $what eq 'dso';}
}if ($disabled{"dynamic-engine"}) {push @{$config{openssl_feature_defines}}, "OPENSSL_NO_DYNAMIC_ENGINE";
} else {push @{$config{openssl_feature_defines}}, "OPENSSL_NO_STATIC_ENGINE";
}# If we use the unified build, collect information from build.info files
my %unified_info = ();my $buildinfo_debug = defined($ENV{CONFIGURE_DEBUG_BUILDINFO});
if ($builder eq "unified") {use Text::Template 1.46;sub cleandir {my $base = shift;my $dir = shift;my $relativeto = shift || ".";$dir = catdir($base,$dir) unless isabsolute($dir);# Make sure the directories we're building in existsmkpath($dir);my $res = abs2rel(absolutedir($dir), rel2abs($relativeto));#print STDERR "DEBUG[cleandir]: $dir , $base => $res\n";return $res;}sub cleanfile {my $base = shift;my $file = shift;my $relativeto = shift || ".";$file = catfile($base,$file) unless isabsolute($file);my $d = dirname($file);my $f = basename($file);# Make sure the directories we're building in existsmkpath($d);my $res = abs2rel(catfile(absolutedir($d), $f), rel2abs($relativeto));#print STDERR "DEBUG[cleanfile]: $d , $f => $res\n";return $res;}# Store the name of the template file we will build the build file from# in %config. This may be useful for the build file itself.my @build_file_template_names =( $builder_platform."-".$config{build_file}.".tmpl",$config{build_file}.".tmpl" );my @build_file_templates = ();# First, look in the user provided directory, if givenif (defined env($local_config_envname)) {@build_file_templates =map {if ($^O eq 'VMS') {# VMS environment variables are logical names,# which can be used as is$local_config_envname . ':' . $_;} else {catfile(env($local_config_envname), $_);}}@build_file_template_names;}# Then, look in our standard directorypush @build_file_templates,( map { cleanfile($srcdir, catfile("Configurations", $_), $blddir) }@build_file_template_names );my $build_file_template;for $_ (@build_file_templates) {$build_file_template = $_;last if -f $build_file_template;$build_file_template = undef;}if (!defined $build_file_template) {die "*** Couldn't find any of:\n", join("\n", @build_file_templates), "\n";}$config{build_file_templates}= [ cleanfile($srcdir, catfile("Configurations", "common0.tmpl"),$blddir),$build_file_template ];my @build_dirs = ( [ ] ); # current directory$config{build_infos} = [ ];# We want to detect configdata.pm in the source tree, so we# don't use it if the build tree is different.my $src_configdata = cleanfile($srcdir, "configdata.pm", $blddir);# Any source file that we recognise is placed in this hash table, with# the list of its intended destinations as value. When everything has# been collected, there's a routine that checks that these source files# exist, or if they are generated, that the generator exists.my %check_exist = ();my %check_generate = ();my %ordinals = ();while (@build_dirs) {my @curd = @{shift @build_dirs};my $sourced = catdir($srcdir, @curd);my $buildd = catdir($blddir, @curd);my $unixdir = join('/', @curd);if (exists $skipdir{$unixdir}) {my $what = $skipdir{$unixdir};push @{$disabled_info{$what}->{skipped}}, catdir(@curd);next;}mkpath($buildd);my $f = 'build.info';# The basic things we're trying to buildmy @programs = ();my @libraries = ();my @modules = ();my @scripts = ();my %sources = ();my %shared_sources = ();my %includes = ();my %defines = ();my %depends = ();my %generate = ();my %imagedocs = ();my %htmldocs = ();my %mandocs = ();# Support for $variablename in build.info files.# Embedded perl code is the ultimate master, still. If its output# contains a dollar sign, it had better be escaped, or it will be# taken for a variable name prefix.my %variables = ();# Variable name syntaxmy $variable_name_re = qr/(?P<VARIABLE>[[:alpha:]][[:alnum:]_]*)/;# Value modifier syntaxesmy $variable_subst_re = qr/\/(?P<RE>(?:\\\/|.)*?)\/(?P<SUBST>.*?)/;# Variable referencemy $variable_simple_re = qr/(?<!\\)\$${variable_name_re}/;my $variable_w_mod_re =qr/(?<!\\)\$\{${variable_name_re}(?P<MOD>(?:\\\/|.)*?)\}/;# Tie it all togethermy $variable_re = qr/${variable_simple_re}|${variable_w_mod_re}/;my $expand_variables = sub {my $value = '';my $value_rest = shift;if ($ENV{CONFIGURE_DEBUG_VARIABLE_EXPAND}) {print STDERR"DEBUG[\$expand_variables] Parsed '$value_rest' ...\n"}while ($value_rest =~ /${variable_re}/) {# We must save important regexp values, because the next# regexp clears themmy $mod = $+{MOD};my $variable_value = $variables{$+{VARIABLE}};$value_rest = $';$value .= $`;# Process modifier expressions, if presentif (defined $mod) {if ($mod =~ /^${variable_subst_re}$/) {my $re = $+{RE};my $subst = $+{SUBST};$variable_value =~ s/\Q$re\E/$subst/g;if ($ENV{CONFIGURE_DEBUG_VARIABLE_EXPAND}) {print STDERR"DEBUG[\$expand_variables] ... and substituted ","'$re' with '$subst'\n";}}}$value .= $variable_value;}if ($ENV{CONFIGURE_DEBUG_VARIABLE_EXPAND}) {print STDERR"DEBUG[\$expand_variables] ... into: '$value$value_rest'\n";}return $value . $value_rest;};# Support for attributes in build.info filesmy %attributes = ();my $handle_attributes = sub {my $attr_str = shift;my $ref = shift;my @goals = @_;return unless defined $attr_str;my @a = tokenize($attr_str, qr|\s*,\s*|);foreach my $a (@a) {my $ac = 1;my $ak = $a;my $av = 1;if ($a =~ m|^(!)?(.*?)\s* = \s*(.*?)$|x) {$ac = ! $1;$ak = $2;$av = $3;}foreach my $g (@goals) {if ($ac) {$$ref->{$g}->{$ak} = $av;} else {delete $$ref->{$g}->{$ak};}}}};# Support for pushing values on multiple indexes of a given hash# array.my $push_to = sub {my $valueref = shift;my $index_str = shift; # May be undef or emptymy $attrref = shift; # May be undefmy $attr_str = shift;my @values = @_;if (defined $index_str) {my @indexes = ( '' );if ($index_str !~ m|^\s*$|) {@indexes = tokenize($index_str);}foreach (@indexes) {push @{$valueref->{$_}}, @values;if (defined $attrref) {$handle_attributes->($attr_str, \$$attrref->{$_},@values);}}} else {push @$valueref, @values;$handle_attributes->($attr_str, $attrref, @values)if defined $attrref;}};if ($buildinfo_debug) {print STDERR "DEBUG: Reading ",catfile($sourced, $f),"\n";}push @{$config{build_infos}}, catfile(abs2rel($sourced, $blddir), $f);my $template =Text::Template->new(TYPE => 'FILE',SOURCE => catfile($sourced, $f),PREPEND => qq{use lib "$FindBin::Bin/util/perl";});die "Something went wrong with $sourced/$f: $!\n" unless $template;my @text =split /^/m,$template->fill_in(HASH => { config => \%config,target => \%target,disabled => \%disabled,withargs => \%withargs,builddir => abs2rel($buildd, $blddir),sourcedir => abs2rel($sourced, $blddir),buildtop => abs2rel($blddir, $blddir),sourcetop => abs2rel($srcdir, $blddir) },DELIMITERS => [ "{-", "-}" ]);# The top item of this stack has the following values# -2 positive already run and we found ELSE (following ELSIF should fail)# -1 positive already run (skip until ENDIF)# 0 negatives so far (if we're at a condition, check it)# 1 last was positive (don't skip lines until next ELSE, ELSIF or ENDIF)# 2 positive ELSE (following ELSIF should fail)my @skip = ();# A few useful generic regexpsmy $index_re = qr/\[\s*(?P<INDEX>(?:\\.|.)*?)\s*\]/;my $cond_re = qr/\[\s*(?P<COND>(?:\\.|.)*?)\s*\]/;my $attribs_re = qr/(?:\{\s*(?P<ATTRIBS>(?:\\.|.)*?)\s*\})?/;my $value_re = qr/(?P<VALUE>.*?)/;collect_information(collect_from_array([ @text ],qr/\\$/ => sub { my $l1 = shift; my $l2 = shift;$l1 =~ s/\\$//; $l1.$l2 }),# Info we're looking forqr/^\s* IF ${cond_re} \s*$/x=> sub {if (! @skip || $skip[$#skip] > 0) {push @skip, !! $expand_variables->($+{COND});} else {push @skip, -1;}},qr/^\s* ELSIF ${cond_re} \s*$/x=> sub { die "ELSIF out of scope" if ! @skip;die "ELSIF following ELSE" if abs($skip[$#skip]) == 2;$skip[$#skip] = -1 if $skip[$#skip] != 0;$skip[$#skip] = !! $expand_variables->($+{COND})if $skip[$#skip] == 0; },qr/^\s* ELSE \s*$/x=> sub { die "ELSE out of scope" if ! @skip;$skip[$#skip] = -2 if $skip[$#skip] != 0;$skip[$#skip] = 2 if $skip[$#skip] == 0; },qr/^\s* ENDIF \s*$/x=> sub { die "ENDIF out of scope" if ! @skip;pop @skip; },qr/^\s* ${variable_re} \s* = \s* ${value_re} \s* $/x=> sub {if (!@skip || $skip[$#skip] > 0) {$variables{$+{VARIABLE}} = $expand_variables->($+{VALUE});}},qr/^\s* SUBDIRS \s* = \s* ${value_re} \s* $/x=> sub {if (!@skip || $skip[$#skip] > 0) {foreach (tokenize($expand_variables->($+{VALUE}))) {push @build_dirs, [ @curd, splitdir($_, 1) ];}}},qr/^\s* PROGRAMS ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\@programs, undef,\$attributes{programs}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* LIBS ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\@libraries, undef,\$attributes{libraries}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* MODULES ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\@modules, undef,\$attributes{modules}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* SCRIPTS ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\@scripts, undef,\$attributes{scripts}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* IMAGEDOCS ${index_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%imagedocs, $expand_variables->($+{INDEX}),undef, undef,tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* HTMLDOCS ${index_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%htmldocs, $expand_variables->($+{INDEX}),undef, undef,tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* MANDOCS ${index_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%mandocs, $expand_variables->($+{INDEX}),undef, undef,tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* SOURCE ${index_re} ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%sources, $expand_variables->($+{INDEX}),\$attributes{sources}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* SHARED_SOURCE ${index_re} ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%shared_sources, $expand_variables->($+{INDEX}),\$attributes{sources}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* INCLUDE ${index_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%includes, $expand_variables->($+{INDEX}),undef, undef,tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* DEFINE ${index_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%defines, $expand_variables->($+{INDEX}),undef, undef,tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* DEPEND ${index_re} ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%depends, $expand_variables->($+{INDEX}),\$attributes{depends}, $+{ATTRIBS},tokenize($expand_variables->($+{VALUE})))if !@skip || $skip[$#skip] > 0; },qr/^\s* GENERATE ${index_re} ${attribs_re} \s* = \s* ${value_re} \s* $/x=> sub { $push_to->(\%generate, $expand_variables->($+{INDEX}),\$attributes{generate}, $+{ATTRIBS},$expand_variables->($+{VALUE}))if !@skip || $skip[$#skip] > 0; },qr/^\s* (?:\#.*)? $/x => sub { },"OTHERWISE" => sub { die "Something wrong with this line:\n$_\nat $sourced/$f" },"BEFORE" => sub {if ($buildinfo_debug) {print STDERR "DEBUG: Parsing ",join(" ", @_),"\n";print STDERR "DEBUG: ... before parsing, skip stack is ",join(" ", map { int($_) } @skip),"\n";}},"AFTER" => sub {if ($buildinfo_debug) {print STDERR "DEBUG: .... after parsing, skip stack is ",join(" ", map { int($_) } @skip),"\n";}},);die "runaway IF?" if (@skip);if (grep { defined $attributes{modules}->{$_}->{engine} } keys %attributesand !$config{dynamic_engines}) {die <<"EOF"
ENGINES can only be used if configured with 'dynamic-engine'.
This is usually a fault in a build.info file.
EOF}{my %infos = ( programs => [ @programs ],libraries => [ @libraries ],modules => [ @modules ],scripts => [ @scripts ] );foreach my $k (keys %infos) {foreach (@{$infos{$k}}) {my $item = cleanfile($buildd, $_, $blddir);$unified_info{$k}->{$item} = 1;# Fix up associated attributes$unified_info{attributes}->{$k}->{$item} =$attributes{$k}->{$_}if defined $attributes{$k}->{$_};}}}# Check that we haven't defined any library as both shared and# explicitly static. That is forbidden.my @doubles = ();foreach (grep /\.a$/, keys %{$unified_info{libraries}}) {(my $l = $_) =~ s/\.a$//;push @doubles, $l if defined $unified_info{libraries}->{$l};}die "these libraries are both explicitly static and shared:\n ",join(" ", @doubles), "\n"if @doubles;foreach (keys %sources) {my $dest = $_;my $ddest = cleanfile($buildd, $_, $blddir);foreach (@{$sources{$dest}}) {my $s = cleanfile($sourced, $_, $blddir);# If it's generated or we simply don't find it in the source# tree, we assume it's in the build tree.if ($s eq $src_configdata || $generate{$_} || ! -f $s) {$s = cleanfile($buildd, $_, $blddir);}my $o = $_;# We recognise C++, C and asm filesif ($s =~ /\.(cc|cpp|c|s|S)$/) {push @{$check_exist{$s}}, $ddest;$o =~ s/\.[csS]$/.o/; # C and assembler$o =~ s/\.(cc|cpp)$/_cc.o/; # C++$o = cleanfile($buildd, $o, $blddir);$unified_info{sources}->{$ddest}->{$o} = -1;$unified_info{sources}->{$o}->{$s} = -1;} elsif ($s =~ /\.rc$/) {# We also recognise resource filespush @{$check_exist{$s}}, $ddest;$o =~ s/\.rc$/.res/; # Resource configuration$o = cleanfile($buildd, $o, $blddir);$unified_info{sources}->{$ddest}->{$o} = -1;$unified_info{sources}->{$o}->{$s} = -1;} else {push @{$check_exist{$s}}, $ddest;$unified_info{sources}->{$ddest}->{$s} = 1;}# Fix up associated attributesif ($o ne $_) {$unified_info{attributes}->{sources}->{$ddest}->{$o} =$unified_info{attributes}->{sources}->{$o}->{$s} =$attributes{sources}->{$dest}->{$_}if defined $attributes{sources}->{$dest}->{$_};} else {$unified_info{attributes}->{sources}->{$ddest}->{$s} =$attributes{sources}->{$dest}->{$_}if defined $attributes{sources}->{$dest}->{$_};}}}foreach (keys %shared_sources) {my $dest = $_;my $ddest = cleanfile($buildd, $_, $blddir);foreach (@{$shared_sources{$dest}}) {my $s = cleanfile($sourced, $_, $blddir);# If it's generated or we simply don't find it in the source# tree, we assume it's in the build tree.if ($s eq $src_configdata || $generate{$_} || ! -f $s) {$s = cleanfile($buildd, $_, $blddir);}my $o = $_;if ($s =~ /\.(cc|cpp|c|s|S)$/) {# We recognise C++, C and asm filespush @{$check_exist{$s}}, $ddest;$o =~ s/\.[csS]$/.o/; # C and assembler$o =~ s/\.(cc|cpp)$/_cc.o/; # C++$o = cleanfile($buildd, $o, $blddir);$unified_info{shared_sources}->{$ddest}->{$o} = -1;$unified_info{sources}->{$o}->{$s} = -1;} elsif ($s =~ /\.rc$/) {# We also recognise resource filespush @{$check_exist{$s}}, $ddest;$o =~ s/\.rc$/.res/; # Resource configuration$o = cleanfile($buildd, $o, $blddir);$unified_info{shared_sources}->{$ddest}->{$o} = -1;$unified_info{sources}->{$o}->{$s} = -1;} elsif ($s =~ /\.ld$/) {# We also recognise linker scripts (or corresponding)# We know they are generated filespush @{$check_exist{$s}}, $ddest;$o = cleanfile($buildd, $_, $blddir);$unified_info{shared_sources}->{$ddest}->{$o} = 1;} else {die "unrecognised source file type for shared library: $s\n";}# Fix up associated attributesif ($o ne $_) {$unified_info{attributes}->{shared_sources}->{$ddest}->{$o} =$unified_info{attributes}->{sources}->{$o}->{$s} =$attributes{sources}->{$dest}->{$_}if defined $attributes{sources}->{$dest}->{$_};} else {$unified_info{attributes}->{shared_sources}->{$ddest}->{$o} =$attributes{sources}->{$dest}->{$_}if defined $attributes{sources}->{$dest}->{$_};}}}foreach (keys %generate) {my $dest = $_;my $ddest = cleanfile($buildd, $_, $blddir);die "more than one generator for $dest: ",join(" ", @{$generate{$_}}),"\n"if scalar @{$generate{$_}} > 1;my @generator = split /\s+/, $generate{$dest}->[0];my $gen = $generator[0];$generator[0] = cleanfile($sourced, $gen, $blddir);# If the generator is itself generated, it's in the build treeif ($generate{$gen} || ! -f $generator[0]) {$generator[0] = cleanfile($buildd, $gen, $blddir);}$check_generate{$ddest}->{$generator[0]}++;$unified_info{generate}->{$ddest} = [ @generator ];# Fix up associated attributes$unified_info{attributes}->{generate}->{$ddest} =$attributes{generate}->{$dest}->{$gen}if defined $attributes{generate}->{$dest}->{$gen};}foreach (keys %depends) {my $dest = $_;my $ddest = $dest;if ($dest =~ /^\|(.*)\|$/) {# Collect the raw target$unified_info{targets}->{$1} = 1;$ddest = $1;} elsif ($dest eq '') {$ddest = '';} else {$ddest = cleanfile($sourced, $dest, $blddir);# If the destination doesn't exist in source, it can only be# a generated file in the build tree.if ($ddest eq $src_configdata || ! -f $ddest) {$ddest = cleanfile($buildd, $dest, $blddir);}}foreach my $f (@{$depends{$dest}}) {# If the dependency destination is generated, dependencies# may have an extra syntax to separate the intended inclusion# directory from the module to be loaded: a | instead of a# / as directory separator.# Do note that this has to be handled in the build file# template as well.# $i = inclusion path in source directory# $i2 = inclusion path in build directory# $m = module path (within the inclusion path)# $i = full module path in source directory# $i2 = full module path in build directorymy $i; my $i2; my $m; my $d; my $d2;if ($unified_info{generate}->{$ddest}&& $f =~ m/^(.*?)\|(.*)$/) {$i = $1;$m = $2;$i = cleanfile($sourced, $i, $blddir);$i2 = cleanfile($buildd, $i, $blddir);$d = cleanfile($sourced, "$i/$m", $blddir);$d2 = cleanfile($buildd, "$i/$m", $blddir);} else {$d = cleanfile($sourced, $f, $blddir);$d2 = cleanfile($buildd, $f, $blddir);}# If we know it's generated, or assume it is because we can't# find it in the source tree, we set file we depend on to be# in the build tree rather than the source tree.if ($d eq $src_configdata|| (grep { $d2 eq $_ }keys %{$unified_info{generate}})|| ! -f $d) {$d = $d2;$i = $i2;}if ($i) {# Put together the computed inclusion dir with the# original module name. Do note that we conserve the# Unixly path syntax for the module path.$d = "$i|$m";}$unified_info{depends}->{$ddest}->{$d} = 1;# Fix up associated attributes$unified_info{attributes}->{depends}->{$ddest}->{$d} =$attributes{depends}->{$dest}->{$f}if defined $attributes{depends}->{$dest}->{$f};}}foreach (keys %includes) {my $dest = $_;my $ddest = cleanfile($sourced, $_, $blddir);# If the destination doesn't exist in source, it can only be# a generated file in the build tree.if ($ddest eq $src_configdata || ! -f $ddest) {$ddest = cleanfile($buildd, $_, $blddir);}foreach (@{$includes{$dest}}) {my $is = cleandir($sourced, $_, $blddir);my $ib = cleandir($buildd, $_, $blddir);push @{$unified_info{includes}->{$ddest}->{source}}, $isunless grep { $_ eq $is } @{$unified_info{includes}->{$ddest}->{source}};push @{$unified_info{includes}->{$ddest}->{build}}, $ibunless grep { $_ eq $ib } @{$unified_info{includes}->{$ddest}->{build}};}}foreach my $dest (keys %defines) {my $ddest;if ($dest ne "") {$ddest = cleanfile($sourced, $dest, $blddir);# If the destination doesn't exist in source, it can only# be a generated file in the build tree.if (! -f $ddest) {$ddest = cleanfile($buildd, $dest, $blddir);}}foreach my $v (@{$defines{$dest}}) {$v =~ m|^([^=]*)(=.*)?$|;die "0 length macro name not permitted\n" if $1 eq "";if ($dest ne "") {die "$1 defined more than once\n"if defined $unified_info{defines}->{$ddest}->{$1};$unified_info{defines}->{$ddest}->{$1} = $2;} else {die "$1 defined more than once\n"if grep { $v eq $_ } @{$config{defines}};push @{$config{defines}}, $v;}}}foreach my $section (keys %imagedocs) {foreach (@{$imagedocs{$section}}) {my $imagedocs = cleanfile($buildd, $_, $blddir);$unified_info{imagedocs}->{$section}->{$imagedocs} = 1;}}foreach my $section (keys %htmldocs) {foreach (@{$htmldocs{$section}}) {my $htmldocs = cleanfile($buildd, $_, $blddir);$unified_info{htmldocs}->{$section}->{$htmldocs} = 1;}}foreach my $section (keys %mandocs) {foreach (@{$mandocs{$section}}) {my $mandocs = cleanfile($buildd, $_, $blddir);$unified_info{mandocs}->{$section}->{$mandocs} = 1;}}}my $ordinals_text = join(', ', sort keys %ordinals);warn <<"EOF" if $ordinals_text;WARNING: ORDINALS were specified for $ordinals_text
They are ignored and should be replaced with a combination of GENERATE,
DEPEND and SHARED_SOURCE.
EOF# Check that each generated file is only generated oncemy $ambiguous_generation = 0;foreach (sort keys %check_generate) {my @generators = sort keys %{$check_generate{$_}};my $generators_txt = join(', ', @generators);if (scalar @generators > 1) {warn "$_ is GENERATEd by more than one generator ($generators_txt)\n";$ambiguous_generation++;}if ($check_generate{$_}->{$generators[0]} > 1) {warn "INFO: $_ has more than one GENERATE declaration (same generator)\n"}}die "There are ambiguous source file generations\n"if $ambiguous_generation > 0;# All given source files should exist, or if generated, their# generator should exist. This loop ensures this is true.my $missing = 0;foreach my $orig (sort keys %check_exist) {foreach my $dest (@{$check_exist{$orig}}) {if ($orig ne $src_configdata) {if ($orig =~ /\.a$/) {# Static library names may be used as sources, so we# need to detect those and give them special treatment.unless (grep { $_ eq $orig }keys %{$unified_info{libraries}}) {warn "$orig is given as source for $dest, but no such library is built\n";$missing++;}} else {# A source may be generated, and its generator may be# generated as well. We therefore loop to dig out the# first generator.my $gen = $orig;while (my @next = keys %{$check_generate{$gen}}) {$gen = $next[0];}if (! -f $gen) {if ($gen ne $orig) {$missing++;warn "$orig is given as source for $dest, but its generator (leading to $gen) is missing\n";} else {$missing++;warn "$orig is given as source for $dest, but is missing\n";}}}}}}die "There are files missing\n" if $missing > 0;# Go through the sources of all libraries and check that the same basename# doesn't appear more than once. Some static library archivers depend on# them being unique.{my $err = 0;foreach my $prod (keys %{$unified_info{libraries}}) {my @prod_sources =map { keys %{$unified_info{sources}->{$_}} }keys %{$unified_info{sources}->{$prod}};my %srccnt = ();# Count how many times a given each source basename# appears for each product.foreach my $src (@prod_sources) {$srccnt{basename $src}++;}foreach my $src (keys %srccnt) {if ((my $cnt = $srccnt{$src}) > 1) {print STDERR "$src appears $cnt times for the product $prod\n";$err++}}}die if $err > 0;}# Massage the result# If we depend on a header file or a perl module, add an inclusion of# its directory to allow smoothe inclusionforeach my $dest (keys %{$unified_info{depends}}) {next if $dest eq "";foreach my $d (keys %{$unified_info{depends}->{$dest}}) {next unless $d =~ /\.(h|pm)$/;# Take into account when a dependency uses the inclusion|module# syntaxmy $i = $d =~ m/\|/ ? $` : dirname($d);my $spot =$d eq "configdata.pm" || defined($unified_info{generate}->{$d})? 'build' : 'source';push @{$unified_info{includes}->{$dest}->{$spot}}, $iunless grep { $_ eq $i } @{$unified_info{includes}->{$dest}->{$spot}};}}# Go through all intermediary files and change their names to something that# reflects what they will be built for. Note that for some source files,# this leads to duplicate object files because they are used multiple times.# the goal is to rename all object files according to this scheme:# {productname}-{midfix}-{origobjname}.[o|res]# the {midfix} is a keyword indicating the type of product, which is mostly# valuable for libraries since they come in two forms.## This also reorganises the {sources} and {shared_sources} so that the# former only contains ALL object files that are supposed to end up in# static libraries and programs, while the latter contains ALL object files# that are supposed to end up in shared libraries and DSOs.# The main reason for having two different source structures is to allow# the same name to be used for the static and the shared variants of a# library.{# Take copies so we don't get interference from added stuffmy %unified_copy = ();foreach (('sources', 'shared_sources')) {$unified_copy{$_} = { %{$unified_info{$_}} }if defined($unified_info{$_});delete $unified_info{$_};}foreach my $prodtype (('programs', 'libraries', 'modules', 'scripts')) {# $intent serves multi purposes:# - give a prefix for the new object files names# - in the case of libraries, rearrange the object files so static# libraries use the 'sources' structure exclusively, while shared# libraries use the 'shared_sources' structure exclusively.my $intent = {programs => { bin => { src => [ 'sources' ],dst => 'sources' } },libraries => { lib => { src => [ 'sources' ],dst => 'sources' },shlib => { prodselect =>sub { grep !/\.a$/, @_ },src => [ 'sources','shared_sources' ],dst => 'shared_sources' } },modules => { dso => { src => [ 'sources' ],dst => 'sources' } },scripts => { script => { src => [ 'sources' ],dst => 'sources' } }} -> {$prodtype};foreach my $kind (keys %$intent) {next if ($intent->{$kind}->{dst} eq 'shared_sources'&& $disabled{shared});my @src = @{$intent->{$kind}->{src}};my $dst = $intent->{$kind}->{dst};my $prodselect = $intent->{$kind}->{prodselect} // sub { @_ };foreach my $prod ($prodselect->(keys %{$unified_info{$prodtype}})) {# %prod_sources has all applicable objects as keys, and# their corresponding sources as valuesmy %prod_sources =map { $_ => [ keys %{$unified_copy{sources}->{$_}} ] }map { keys %{$unified_copy{$_}->{$prod}} }@src;foreach (keys %prod_sources) {# Only affect object files and resource files,# the others simply get a new value# (+1 instead of -1)if ($_ =~ /\.(o|res)$/) {(my $prodname = $prod) =~ s|\.a$||;my $newobj =catfile(dirname($_),basename($prodname). '-' . $kind. '-' . basename($_));$unified_info{$dst}->{$prod}->{$newobj} = 1;foreach my $src (@{$prod_sources{$_}}) {$unified_info{sources}->{$newobj}->{$src} = 1;# Adjust source attributesmy $attrs = $unified_info{attributes}->{sources};if (defined $attrs->{$prod}&& defined $attrs->{$prod}->{$_}) {$attrs->{$prod}->{$newobj} =$attrs->{$prod}->{$_};delete $attrs->{$prod}->{$_};}foreach my $objsrc (keys %{$attrs->{$_} // {}}) {$attrs->{$newobj}->{$objsrc} =$attrs->{$_}->{$objsrc};delete $attrs->{$_}->{$objsrc};}}# Adjust dependenciesforeach my $deps (keys %{$unified_info{depends}->{$_}}) {$unified_info{depends}->{$_}->{$deps} = -1;$unified_info{depends}->{$newobj}->{$deps} = 1;}# Adjust includesforeach my $k (('source', 'build')) {next unlessdefined($unified_info{includes}->{$_}->{$k});my @incs = @{$unified_info{includes}->{$_}->{$k}};$unified_info{includes}->{$newobj}->{$k} = [ @incs ];}} else {$unified_info{$dst}->{$prod}->{$_} = 1;}}}}}}# At this point, we have a number of sources with the value -1. They# aren't part of the local build and are probably meant for a different# platform, and can therefore be cleaned away. That happens when making# %unified_info more efficient below.### Make unified_info a bit more efficient# One level structuresforeach (("programs", "libraries", "modules", "scripts", "targets")) {$unified_info{$_} = [ sort keys %{$unified_info{$_}} ];}# Two level structuresforeach my $l1 (("sources", "shared_sources", "ldadd", "depends","imagedocs", "htmldocs", "mandocs")) {foreach my $l2 (sort keys %{$unified_info{$l1}}) {my @items =sortgrep { $unified_info{$l1}->{$l2}->{$_} > 0 }keys %{$unified_info{$l1}->{$l2}};if (@items) {$unified_info{$l1}->{$l2} = [ @items ];} else {delete $unified_info{$l1}->{$l2};}}}# Definesforeach my $dest (sort keys %{$unified_info{defines}}) {$unified_info{defines}->{$dest}= [ map { $_.$unified_info{defines}->{$dest}->{$_} }sort keys %{$unified_info{defines}->{$dest}} ];}# Includesforeach my $dest (sort keys %{$unified_info{includes}}) {if (defined($unified_info{includes}->{$dest}->{build})) {my @source_includes = ();@source_includes = ( @{$unified_info{includes}->{$dest}->{source}} )if defined($unified_info{includes}->{$dest}->{source});$unified_info{includes}->{$dest} =[ @{$unified_info{includes}->{$dest}->{build}} ];foreach my $inc (@source_includes) {push @{$unified_info{includes}->{$dest}}, $incunless grep { $_ eq $inc } @{$unified_info{includes}->{$dest}};}} elsif (defined($unified_info{includes}->{$dest}->{source})) {$unified_info{includes}->{$dest} =[ @{$unified_info{includes}->{$dest}->{source}} ];} else {delete $unified_info{includes}->{$dest};}}# For convenience collect information regarding directories where# files are generated, those generated files and the end product# they end up in where applicable. Then, add build rules for those# directoriesmy %loopinfo = ( "lib" => [ @{$unified_info{libraries}} ],"dso" => [ @{$unified_info{modules}} ],"bin" => [ @{$unified_info{programs}} ],"script" => [ @{$unified_info{scripts}} ],"docs" => [ (map { @{$unified_info{imagedocs}->{$_} // []} }keys %{$unified_info{imagedocs} // {}}),(map { @{$unified_info{htmldocs}->{$_} // []} }keys %{$unified_info{htmldocs} // {}}),(map { @{$unified_info{mandocs}->{$_} // []} }keys %{$unified_info{mandocs} // {}}) ] );foreach my $type (sort keys %loopinfo) {foreach my $product (@{$loopinfo{$type}}) {my %dirs = ();my $pd = dirname($product);foreach (@{$unified_info{sources}->{$product} // []},@{$unified_info{shared_sources}->{$product} // []}) {my $d = dirname($_);# We don't want to create targets for source directories# when building out of sourcenext if ($config{sourcedir} ne $config{builddir}&& $d =~ m|^\Q$config{sourcedir}\E|);# We already have a "test" target, and the current directory# is just silly to make a target fornext if $d eq "test" || $d eq ".";$dirs{$d} = 1;push @{$unified_info{dirinfo}->{$d}->{deps}}, $_if $d ne $pd;}foreach (sort keys %dirs) {push @{$unified_info{dirinfo}->{$_}->{products}->{$type}},$product;}}}
}# For the schemes that need it, we provide the old *_obj configs
# from the *_asm_obj ones
foreach (grep /_(asm|aux)_src$/, keys %target) {my $src = $_;(my $obj = $_) =~ s/_(asm|aux)_src$/_obj/;$target{$obj} = $target{$src};$target{$obj} =~ s/\.[csS]\b/.o/g; # C and assembler$target{$obj} =~ s/\.(cc|cpp)\b/_cc.o/g; # C++
}# Write down our configuration where it fits #########################my %template_vars = (config => \%config,target => \%target,disablables => \@disablables,disablables_int => \@disablables_int,disabled => \%disabled,withargs => \%withargs,unified_info => \%unified_info,tls => \@tls,dtls => \@dtls,makevars => [ sort keys %user ],disabled_info => \%disabled_info,user_crossable => \@user_crossable,
);
my $configdata_outname = 'configdata.pm';
open CONFIGDATA, ">$configdata_outname.new"or die "Trying to create $configdata_outname.new: $!";
my $configdata_tmplname = cleanfile($srcdir, "configdata.pm.in", $blddir);
my $configdata_tmpl =OpenSSL::Template->new(TYPE => 'FILE', SOURCE => $configdata_tmplname);
$configdata_tmpl->fill_in(FILENAME => $configdata_tmplname,OUTPUT => \*CONFIGDATA,HASH => { %template_vars,autowarntext => ['WARNING: do not edit!',"Generated by Configure from $configdata_tmplname",] }
) or die $Text::Template::ERROR;
close CONFIGDATA;rename "$configdata_outname.new", $configdata_outname;
if ($builder_platform eq 'unix') {my $mode = (0755 & ~umask);chmod $mode, 'configdata.pm'or warn sprintf("WARNING: Couldn't change mode for 'configdata.pm' to 0%03o: %s\n",$mode,$!);
}
print "Created $configdata_outname\n";print "Running $configdata_outname\n";
my $perlcmd = (quotify("maybeshell", $config{PERL}))[0];
my $cmd = "$perlcmd $configdata_outname";
#print STDERR "DEBUG[run_dofile]: \$cmd = $cmd\n";
system($cmd);
exit 1 if $? != 0;$SIG{__DIE__} = $orig_death_handler;print <<"EOF" if ($disabled{threads} eq "unavailable");The library could not be configured for supporting multi-threaded
applications as the compiler options required on this system are not known.
See file INSTALL.md for details if you need multi-threading.
EOFprint <<"EOF" if ($no_shared_warn);The options 'shared', 'pic' and 'dynamic-engine' aren't supported on this
platform, so we will pretend you gave the option 'no-pic', which also disables
'shared' and 'dynamic-engine'. If you know how to implement shared libraries
or position independent code, please let us know (but please first make sure
you have tried with a current version of OpenSSL).
EOFprint $banner;exit(0);######################################################################
#
# Helpers and utility functions
## Death handler, to print a helpful message in case of failure #######
#
sub death_handler {die @_ if $^S; # To prevent the added message in eval blocksmy $build_file = $config{build_file} // "build file";my @message = ( <<"_____", @_ );Failure! $build_file wasn't produced.
Please read INSTALL.md and associated NOTES-* files. You may also have to
look over your available compiler tool chain or change your configuration._____# Dying is terminal, so it's ok to reset the signal handler here.$SIG{__DIE__} = $orig_death_handler;die @message;
}# Configuration file reading ########################################## Note: All of the helper functions are for lazy evaluation. They all
# return a CODE ref, which will return the intended value when evaluated.
# Thus, whenever there's mention of a returned value, it's about that
# intended value.# Helper function to implement conditional value variants, with a default
# plus additional values based on the value of $config{build_type}.
# Arguments are given in hash table form:
#
# picker(default => "Basic string: ",
# debug => "debug",
# release => "release")
#
# When configuring with --debug, the resulting string will be
# "Basic string: debug", and when not, it will be "Basic string: release"
#
# This can be used to create variants of sets of flags according to the
# build type:
#
# cflags => picker(default => "-Wall",
# debug => "-g -O0",
# release => "-O3")
#
sub picker {my %opts = @_;return sub { add($opts{default} || (),$opts{$config{build_type}} || ())->(); }
}# Helper function to combine several values of different types into one.
# This is useful if you want to combine a string with the result of a
# lazy function, such as:
#
# cflags => combine("-Wall", sub { $disabled{zlib} ? () : "-DZLIB" })
#
sub combine {my @stuff = @_;return sub { add(@stuff)->(); }
}# Helper function to implement conditional values depending on the value
# of $disabled{threads}. Can be used as follows:
#
# cflags => combine("-Wall", threads("-pthread"))
#
sub threads {my @flags = @_;return sub { add($disabled{threads} ? () : @flags)->(); }
}sub shared {my @flags = @_;return sub { add($disabled{shared} ? () : @flags)->(); }
}our $add_called = 0;
# Helper function to implement adding values to already existing configuration
# values. It handles elements that are ARRAYs, CODEs and scalars
sub _add {my $separator = shift;# If there's any ARRAY in the collection of values OR the separator# is undef, we will return an ARRAY of combined values, otherwise a# string of joined values with $separator as the separator.my $found_array = !defined($separator);my @values =map {my $res = $_;while (ref($res) eq "CODE") {$res = $res->();}if (defined($res)) {if (ref($res) eq "ARRAY") {$found_array = 1;@$res;} else {$res;}} else {();}} (@_);$add_called = 1;if ($found_array) {[ @values ];} else {join($separator, grep { defined($_) && $_ ne "" } @values);}
}
sub add_before {my $separator = " ";if (ref($_[$#_]) eq "HASH") {my $opts = pop;$separator = $opts->{separator};}my @x = @_;sub { _add($separator, @x, @_) };
}
sub add {my $separator = " ";if (ref($_[$#_]) eq "HASH") {my $opts = pop;$separator = $opts->{separator};}my @x = @_;sub { _add($separator, @_, @x) };
}sub read_eval_file {my $fname = shift;my $content;my @result;open F, "< $fname" or die "Can't open '$fname': $!\n";{undef local $/;$content = <F>;}close F;{local $@;@result = ( eval $content );warn $@ if $@;}return wantarray ? @result : $result[0];
}# configuration reader, evaluates the input file as a perl script and expects
# it to fill %targets with target configurations. Those are then added to
# %table.
sub read_config {my $fname = shift;my %targets;{# Protect certain tables from tamperinglocal %table = ();%targets = read_eval_file($fname);}my %preexisting = ();foreach (sort keys %targets) {$preexisting{$_} = 1 if $table{$_};}die <<"EOF",
The following config targets from $fname
shadow pre-existing config targets with the same name:
EOFmap { " $_\n" } sort keys %preexistingif %preexisting;# For each target, check that it's configured with a hash table.foreach (keys %targets) {if (ref($targets{$_}) ne "HASH") {if (ref($targets{$_}) eq "") {warn "Deprecated target configuration for $_, ignoring...\n";} else {warn "Misconfigured target configuration for $_ (should be a hash table), ignoring...\n";}delete $targets{$_};} else {$targets{$_}->{_conf_fname_int} = add([ $fname ]);}}%table = (%table, %targets);}# configuration resolver. Will only resolve all the lazy evaluation
# codeblocks for the chosen target and all those it inherits from,
# recursively
sub resolve_config {my $target = shift;my @breadcrumbs = @_;# my $extra_checks = defined($ENV{CONFIGURE_EXTRA_CHECKS});if (grep { $_ eq $target } @breadcrumbs) {die "inherit_from loop! target backtrace:\n ",$target,"\n ",join("\n ", @breadcrumbs),"\n";}if (!defined($table{$target})) {warn "Warning! target $target doesn't exist!\n";return ();}# Recurse through all inheritances. They will be resolved on the# fly, so when this operation is done, they will all just be a# bunch of attributes with string values.# What we get here, though, are keys with references to lists of# the combined values of them all. We will deal with lists after# this stage is done.my %combined_inheritance = ();if ($table{$target}->{inherit_from}) {my @inherit_from =map { ref($_) eq "CODE" ? $_->() : $_ } @{$table{$target}->{inherit_from}};foreach (@inherit_from) {my %inherited_config = resolve_config($_, $target, @breadcrumbs);# 'template' is a marker that's considered private to# the config that had it.delete $inherited_config{template};foreach (keys %inherited_config) {if (!$combined_inheritance{$_}) {$combined_inheritance{$_} = [];}push @{$combined_inheritance{$_}}, $inherited_config{$_};}}}# We won't need inherit_from in this target any more, since we've# resolved all the inheritances that lead to thisdelete $table{$target}->{inherit_from};# Now is the time to deal with those lists. Here's the place to# decide what shall be done with those lists, all based on the# values of the target we're currently dealing with.# - If a value is a coderef, it will be executed with the list of# inherited values as arguments.# - If the corresponding key doesn't have a value at all or is the# empty string, the inherited value list will be run through the# default combiner (below), and the result becomes this target's# value.# - Otherwise, this target's value is assumed to be a string that# will simply override the inherited list of values.my $default_combiner = add();my %all_keys =map { $_ => 1 } (keys %combined_inheritance,keys %{$table{$target}});sub process_values {my $object = shift;my $inherited = shift; # Always a [ list ]my $target = shift;my $entry = shift;$add_called = 0;while(ref($object) eq "CODE") {$object = $object->(@$inherited);}if (!defined($object)) {return ();}elsif (ref($object) eq "ARRAY") {local $add_called; # To make sure recursive calls don't affect itreturn [ map { process_values($_, $inherited, $target, $entry) }@$object ];} elsif (ref($object) eq "") {return $object;} else {die "cannot handle reference type ",ref($object)," found in target ",$target," -> ",$entry,"\n";}}foreach my $key (sort keys %all_keys) {my $previous = $combined_inheritance{$key};# Current target doesn't have a value for the current key?# Assign it the default combiner, the rest of this loop body# will handle it just like any other coderef.if (!exists $table{$target}->{$key}) {$table{$target}->{$key} = $default_combiner;}$table{$target}->{$key} = process_values($table{$target}->{$key},$combined_inheritance{$key},$target, $key);unless(defined($table{$target}->{$key})) {delete $table{$target}->{$key};}
# if ($extra_checks &&
# $previous && !($add_called || $previous ~~ $table{$target}->{$key})) {
# warn "$key got replaced in $target\n";
# }}# Finally done, return the result.return %{$table{$target}};
}sub usage{print STDERR $usage;print STDERR "\npick os/compiler from:\n";my $j=0;my $i;my $k=0;foreach $i (sort keys %table){next if $table{$i}->{template};next if $i =~ /^debug/;$k += length($i) + 1;if ($k > 78){print STDERR "\n";$k=length($i);}print STDERR $i . " ";}foreach $i (sort keys %table){next if $table{$i}->{template};next if $i !~ /^debug/;$k += length($i) + 1;if ($k > 78){print STDERR "\n";$k=length($i);}print STDERR $i . " ";}exit(1);}sub compiler_predefined {state %predefined;my $cc = shift;return () if $^O eq 'VMS';die 'compiler_predefined called without a compiler command'unless $cc;if (! $predefined{$cc}) {$predefined{$cc} = {};# collect compiler pre-defines from gcc or gcc-alike...open(PIPE, "$cc -dM -E -x c /dev/null 2>&1 |");while (my $l = <PIPE>) {$l =~ m/^#define\s+(\w+(?:\(\w+\))?)(?:\s+(.+))?/ or last;$predefined{$cc}->{$1} = $2 // '';}close(PIPE);}return %{$predefined{$cc}};
}sub which
{my ($name)=@_;if (eval { require IPC::Cmd; 1; }) {IPC::Cmd->import();return scalar IPC::Cmd::can_run($name);} else {# if there is $directories component in splitpath,# then it's not something to test with $PATH...return $name if (File::Spec->splitpath($name))[1];foreach (File::Spec->path()) {my $fullpath = catfile($_, "$name$target{exe_extension}");if (-f $fullpath and -x $fullpath) {return $fullpath;}}}
}sub env
{my $name = shift;my %opts = @_;unless ($opts{cacheonly}) {# Note that if $ENV{$name} doesn't exist or is undefined,# $config{perlenv}->{$name} will be created with the value# undef. This is intentional.$config{perlenv}->{$name} = $ENV{$name}if ! exists $config{perlenv}->{$name};}return $config{perlenv}->{$name};
}# Configuration printer ##############################################sub print_table_entry
{local $now_printing = shift;my %target = resolve_config($now_printing);my $type = shift;# Don't print the templatesreturn if $target{template};my @sequence = ("sys_id","cpp","cppflags","defines","includes","cc","cflags","ld","lflags","loutflag","ex_libs","bn_ops","enable","disable","poly1035_asm_src","thread_scheme","perlasm_scheme","dso_scheme","shared_target","shared_cflag","shared_defines","shared_ldflag","shared_rcflag","shared_extension","dso_extension","obj_extension","exe_extension","ranlib","ar","arflags","aroutflag","rc","rcflags","rcoutflag","mt","mtflags","mtinflag","mtoutflag","multilib","build_scheme",);if ($type eq "TABLE") {print "\n";print "*** $now_printing\n";foreach (@sequence) {if (ref($target{$_}) eq "ARRAY") {printf "\$%-12s = %s\n", $_, join(" ", @{$target{$_}});} else {printf "\$%-12s = %s\n", $_, $target{$_};}}} elsif ($type eq "HASH") {my $largest =length((sort { length($a) <=> length($b) } @sequence)[-1]);print " '$now_printing' => {\n";foreach (@sequence) {if ($target{$_}) {if (ref($target{$_}) eq "ARRAY") {print " '",$_,"'"," " x ($largest - length($_))," => [ ",join(", ", map { "'$_'" } @{$target{$_}})," ],\n";} else {print " '",$_,"'"," " x ($largest - length($_))," => '",$target{$_},"',\n";}}}print " },\n";}
}# Utility routines #################################################### On VMS, if the given file is a logical name, File::Spec::Functions
# will consider it an absolute path. There are cases when we want a
# purely syntactic check without checking the environment.
sub isabsolute {my $file = shift;# On non-platforms, we just use file_name_is_absolute().return file_name_is_absolute($file) unless $^O eq "VMS";# If the file spec includes a device or a directory spec,# file_name_is_absolute() is perfectly safe.return file_name_is_absolute($file) if $file =~ m|[:\[]|;# Here, we know the given file spec isn't absolutereturn 0;
}# Makes a directory absolute and cleans out /../ in paths like foo/../bar
# On some platforms, this uses rel2abs(), while on others, realpath() is used.
# realpath() requires that at least all path components except the last is an
# existing directory. On VMS, the last component of the directory spec must
# exist.
sub absolutedir {my $dir = shift;# realpath() is quite buggy on VMS. It uses LIB$FID_TO_NAME, which# will return the volume name for the device, no matter what. Also,# it will return an incorrect directory spec if the argument is a# directory that doesn't exist.if ($^O eq "VMS") {return rel2abs($dir);}# We use realpath() on Unix, since no other will properly clean out# a directory spec.use Cwd qw/realpath/;return realpath($dir);
}# Check if all paths are one and the same, using stat. They must both exist
# We need this for the cases when File::Spec doesn't detect case insensitivity
# (File::Spec::Unix assumes case sensitivity)
sub samedir {die "samedir expects two arguments\n" unless scalar @_ == 2;my @stat0 = stat($_[0]); # First argumentmy @stat1 = stat($_[1]); # Second argumentdie "Couldn't stat $_[0]" unless @stat0;die "Couldn't stat $_[1]" unless @stat1;# Compare device numberreturn 0 unless ($stat0[0] == $stat1[0]);# Compare "inode". The perl manual recommends comparing as# string rather than as number.return 0 unless ($stat0[1] eq $stat1[1]);return 1; # All the same
}sub quotify {my %processors = (perl => sub { my $x = shift;$x =~ s/([\\\$\@"])/\\$1/g;return '"'.$x.'"'; },maybeshell => sub { my $x = shift;(my $y = $x) =~ s/([\\\"])/\\$1/g;if ($x ne $y || $x =~ m|\s|) {return '"'.$y.'"';} else {return $x;}},);my $for = shift;my $processor =defined($processors{$for}) ? $processors{$for} : sub { shift; };return map { $processor->($_); } @_;
}# collect_from_file($filename, $line_concat_cond_re, $line_concat)
# $filename is a file name to read from
# $line_concat_cond_re is a regexp detecting a line continuation ending
# $line_concat is a CODEref that takes care of concatenating two lines
sub collect_from_file {my $filename = shift;my $line_concat_cond_re = shift;my $line_concat = shift;open my $fh, $filename || die "unable to read $filename: $!\n";return sub {my $saved_line = "";$_ = "";while (<$fh>) {s|\R$||;if (defined $line_concat) {$_ = $line_concat->($saved_line, $_);$saved_line = "";}if (defined $line_concat_cond_re && /$line_concat_cond_re/) {$saved_line = $_;next;}return $_;}die "$filename ending with continuation line\n" if $_;close $fh;return undef;}
}# collect_from_array($array, $line_concat_cond_re, $line_concat)
# $array is an ARRAYref of lines
# $line_concat_cond_re is a regexp detecting a line continuation ending
# $line_concat is a CODEref that takes care of concatenating two lines
sub collect_from_array {my $array = shift;my $line_concat_cond_re = shift;my $line_concat = shift;my @array = (@$array);return sub {my $saved_line = "";$_ = "";while (defined($_ = shift @array)) {s|\R$||;if (defined $line_concat) {$_ = $line_concat->($saved_line, $_);$saved_line = "";}if (defined $line_concat_cond_re && /$line_concat_cond_re/) {$saved_line = $_;next;}return $_;}die "input text ending with continuation line\n" if $_;return undef;}
}# collect_information($lineiterator, $line_continue, $regexp => $CODEref, ...)
# $lineiterator is a CODEref that delivers one line at a time.
# All following arguments are regex/CODEref pairs, where the regexp detects a
# line and the CODEref does something with the result of the regexp.
sub collect_information {my $lineiterator = shift;my %collectors = @_;while(defined($_ = $lineiterator->())) {s|\R$||;my $found = 0;if ($collectors{"BEFORE"}) {$collectors{"BEFORE"}->($_);}foreach my $re (keys %collectors) {if ($re !~ /^OTHERWISE|BEFORE|AFTER$/ && /$re/) {$collectors{$re}->($lineiterator);$found = 1;};}if ($collectors{"OTHERWISE"}) {$collectors{"OTHERWISE"}->($lineiterator, $_)unless $found || !defined $collectors{"OTHERWISE"};}if ($collectors{"AFTER"}) {$collectors{"AFTER"}->($_);}}
}# tokenize($line)
# tokenize($line,$separator)
# $line is a line of text to split up into tokens
# $separator [optional] is a regular expression that separates the tokens,
# the default being spaces. Do not use quotes of any kind as separators,
# that will give undefined results.
# Returns a list of tokens.
#
# Tokens are divided by separator (spaces by default). If the tokens include
# the separators, they have to be quoted with single or double quotes.
# Double quotes inside a double quoted token must be escaped. Escaping is done
# with backslash.
# Basically, the same quoting rules apply for " and ' as in any
# Unix shell.
sub tokenize {my $line = my $debug_line = shift;my $separator = shift // qr|\s+|;my @result = ();if ($ENV{CONFIGURE_DEBUG_TOKENIZE}) {print STDERR "DEBUG[tokenize]: \$separator = $separator\n";}while ($line =~ s|^${separator}||, $line ne "") {my $token = "";again:$line =~ m/^(.*?)(${separator}|"|'|$)/;$token .= $1;$line = $2.$';if ($line =~ m/^"((?:[^"\\]+|\\.)*)"/) {$token .= $1;$line = $';goto again;} elsif ($line =~ m/^'([^']*)'/) {$token .= $1;$line = $';goto again;}push @result, $token;}if ($ENV{CONFIGURE_DEBUG_TOKENIZE}) {print STDERR "DEBUG[tokenize]: Parsed '$debug_line' into:\n";print STDERR "DEBUG[tokenize]: ('", join("', '", @result), "')\n";}return @result;
}
相关文章:

使用 LLVM clang C/C++ 编译器编译 OpenSSL 3.X库
1、下载 OpenSSL 3.X 库的源代码放到待编译目录 2、解压并接入 OpenSSL 3.X 库源码的根目录 3、复制 ./Configure 一个取名为 ./Configure-clang 4、修改 ./Configure-clang 找到配置段: CC CXX CPP LD 把它们改成 CC > "/usr/bin/clang-…...

【信息安全】hydra爆破工具的使用方法
hydra简介 hydra又名九头蛇,与burp常规的爆破模块不同,hydra爆破的范围更加广泛,可以爆破远程桌面连接,数据库这类的密码。他在kali系统中自带。 参数说明 -l 指定用户名 -L 指定用户名字典文件 -p 指定密码 -P 指…...

uniapp中uview组件库丰富的CountTo 数字滚动使用方法
目录 #平台差异说明 #基本使用 #设置滚动相关参数 #是否显示小数位 #千分位分隔符 #滚动执行的时机 #API #Props #Methods #Event 该组件一般用于需要滚动数字到某一个值的场景,目标要求是一个递增的值。 注意 如果给组件的父元素设置text-align: cente…...

inflate流程分析
一.inflate的三参数重载方法else里面逻辑 我们先看到setContentView里面的inflate的调用链: public View inflate(LayoutRes int resource, Nullable ViewGroup root) {return inflate(resource, root, root ! null);}public View inflate(LayoutRes int resource…...

数据挖掘实战-基于机器学习的电商文本分类模型
🤵♂️ 个人主页:艾派森的个人主页 ✍🏻作者简介:Python学习者 🐋 希望大家多多支持,我们一起进步!😄 如果文章对你有帮助的话, 欢迎评论 💬点赞Ǵ…...

第8章-第4节-Java中字节流的缓冲流
1、缓冲流:属于高级IO流,并不能直接读写数据,需要依赖于基础流。缓冲流的目的是为了提高文件的读写效率?那么是如何提高文件的读写效率的呢? 在内存中设置一个缓冲区,缓冲区的默认大小是8192字节ÿ…...

NULL是什么?
NULL是一个编程术语,通常用于表示一个空值或无效值。在很多编程语言中,NULL用于表示一个变量或指针不引用任何有效的对象或内存位置。 NULL可以看作是一个特殊的值,表示缺少有效的数据或引用。当一个变量被赋予NULL值时,它表示该变…...

FreeRTOS 基础知识
这个基础知识也是非常重要的,那我们要学好 FreeRTOS,这些都是必不可少的。 那么就来看一下本节有哪些内容: 首先呢就是介绍一下什么是任务调度器。接着呢就是任务它拥有哪一些状态了。那这里的内容不多,但是呢都是非常重要的。 …...

【野火i.MX6NULL开发板】挂载 NFS 网络文件系统
0、前言 参考资料: (误人子弟)《野火 Linux 基础与应用开发实战指南基于 i.MX6ULL 系列》PDF 第22章 参考视频:(成功) https://www.bilibili.com/video/BV1JK4y1t7io?p26&vd_sourcefb8dcae0aee3f1aab…...

在JavaScript中,Object.assign()方法或展开语法(...)来合并对象,Object.freeze()方法来冻结对象,防止对象被修改
文章目录 一、Object.freeze()方法来冻结对象,防止对象被修改1、基本使用2、冻结数组2.1、浅冻结2.1、深冻结 3、应用场景4、Vue中使用Object.freeze 二、Object.assign()方法或展开语法(...)来合并对象1、Object.assign()1.1、语法1.2、参数…...

池化、线性、激活函数层
一、池化层 池化运算是深度学习中常用的一种操作,它可以对输入的特征图进行降采样,从而减少特征图的尺寸和参数数量。 池化运算的主要目的是通过“收集”和“总结”输入特征图的信息来提取出主要特征,并且减少对细节的敏感性。在池化运算中…...

ES-极客学习第二部分ES 入门
基本概念 索引、文档、节点、分片和API json 文档 文档的元数据 需要通过Kibana导入Sample Data的电商数据。具体参考“2.2节-Kibana的安装与界面快速浏览” 索引 kibana 管理ES索引 在系统中找到kibana配置文件(我这里是etc/kibana/kibana.yml) vim /…...

Nodejs软件安装
Nodejs软件安装 一、简介 Node.js 是一个基于 Chrome V8 引擎的 JavaScript 运行环境。 官网:http://nodejs.cn/api/ 我们关注于 node.js 的 npm 功能,NPM 是随同 NodeJS 一起安装的包管理工具,JavaScript-NPM,Java-Maven&…...

Photoshop 2024 (PS2024) v25 直装版 支持win/mac版
Photoshop 2024 提供了多种创意工具,如画笔、铅笔、涂鸦和渐变等,用户可以通过这些工具来创建独特和令人印象深刻的设计效果。增强的云同步:通过 Adobe Creative Cloud,用户可以方便地将他们的工作从一个设备无缝同步到另一个设备…...

ChatGPT绘画生成软件MidTool:智能艺术的新纪元
在人工智能的黄金时代,创新技术不断涌现,改变着我们的生活和工作方式。其中,ChatGPT绘画生成软件MidTool无疑是这一变革浪潮中的佼佼者。它不仅是一个软件,更是一位艺术家,一位智能助手,它的出现预示着智能…...

linux安装MySQL5.7(安装、开机自启、定时备份)
一、安装步骤 我喜欢安装在/usr/local/mysql目录下 #切换目录 cd /usr/local/ #下载文件 wget https://dev.mysql.com/get/Downloads/MySQL-5.7/mysql-5.7.38-linux-glibc2.12-x86_64.tar.gz #解压文件 tar -zxvf mysql-5.7.38-linux-glibc2.12-x86_64.tar.gz -C /usr/local …...

openGauss学习笔记-195 openGauss 数据库运维-常见故障定位案例-分析查询语句运行状态
文章目录 openGauss学习笔记-195 openGauss 数据库运维-常见故障定位案例-分析查询语句运行状态195.1 分析查询语句运行状态195.1.1 问题现象195.1.2 处理办法 openGauss学习笔记-195 openGauss 数据库运维-常见故障定位案例-分析查询语句运行状态 195.1 分析查询语句运行状态…...

Oracle篇—实例中和name相关参数的区别和作用
☘️博主介绍☘️: ✨又是一天没白过,我是奈斯,DBA一名✨ ✌✌️擅长Oracle、MySQL、SQLserver、Linux,也在积极的扩展IT方向的其他知识面✌✌️ ❣️❣️❣️大佬们都喜欢静静的看文章,并且也会默默的点赞收藏加关注❣…...

python + selenium 初步实现数据驱动
如果在进行自动化测试的时候将测试数据写在代码中,若测试数据有变,不利于数据的修改和维护。但可以尝试通过将测试数据放到excel文档中来实现测试数据的管理。 示例:本次涉及的项目使用的12306 selenium 重构------三层架构 excel文件数据如…...

数字孪生+可视化技术 构建智慧新能源汽车充电站监管平台
前言 充电基础设施为电动汽车提供充换电服务,是重要的交通能源融合类基础设施。近年来,随着新能源汽车产业快速发展,我国充电基础设施持续增长,已建成世界上数量最多、服务范围最广、品种类型最全的充电基础设施体系。着眼未来新…...

微信小程序开发学习笔记《11》导航传参
微信小程序开发学习笔记《11》导航传参 博主正在学习微信小程序开发,希望记录自己学习过程同时与广大网友共同学习讨论。导航传参 官方文档 一、声明式导航传参 navigator组件的url属性用来指定将要跳转到的页面的路径。同时,路径的后面还可以携带参数…...

BikeDNA(七)外在分析:OSM 与参考数据的比较1
BikeDNA(七)外在分析:OSM 与参考数据的比较1 该笔记本将提供的参考自行车基础设施数据集与同一区域的 OSM 数据进行所谓的外部质量评估进行比较。 为了运行这部分分析,必须有一个参考数据集可用于比较。 该分析基于将参考数据集…...

KY43 全排列
全排列板子 ti #include<bits/stdc.h>using namespace std;string s; map<string, int>mp;void swap(char &a, char &b){char em a;a b;b em; }void dfs(int n){ //将s[n~l]的全排列转化成s[n]s[n1~l]的全排列 if(n s.length()){mp[s] 1;return ;}f…...

UltraScale 和 UltraScale+ 生成已加密文件和已经过身份验证的文件
注释 :如需了解更多信息,请参阅《使用加密和身份验证确保 UltraScale/UltraScale FPGA 比特流的安全》 (XAPP1267)。 要生成加密比特流,请在 Vivado IDE 中打开已实现的设计。在主工具栏中,依次选择“Flow” → “Bitstream Setti…...

2023年全国职业院校技能大赛软件测试赛题—单元测试卷②
单元测试 一、任务要求 题目1:任意输入2个正整数值分别存入x、y中,据此完成下述分析:若x≤0或y≤0,则提示:“输入不符合要求。”;若2值相同,则提示“可以构建圆形或正方形”;若2<…...

极兔单号查快递,极兔快递单号查询,筛选出途经指定城市的单号
随着电商的繁荣,快递单号已经成为我们生活中的一部分。然而,面对海量的快递信息,如何快速、准确地筛选出我们需要的单号,变成了许多人的痛点。今天,我要为你介绍一款强大的工具——快递批量查询高手,让你的…...

[redis] redis高可用之持久化
一、Redis 高可用的相关知识 1.1 什么是高可用 在web服务器中,高可用是指服务器可以正常访问的时间,衡量的标准是在多长时间内可以提供正常服务(99.9%、99.99%、99.999%等等)。 但是在Redis语境中,高可用的含义似乎要宽泛一些,…...

云原生 微服务 restapi devops相关的一些概念说明(持续更新中)
云原生: 定义 云原生是一种构建和运行应用程序的方法,是一套技术体系和方法论。它是一种在云计算环境中构建、部署和管理现代应用程序的软件方法。云原生应用程序是基于微服务架构的,采用开源堆栈(K8SDocker)进行容器…...

初学unity学习七天,经验收获总结
初学unity七天,经验收获总结 学习就是认识新观念和新想法的过程。 假如人们始终以同一种思维方式来考虑问题的话,那么始终只会得到同样的结果。 因为我对你讲述的许多内容是你以前从未接触过的,所以我建议你,在你还没有做之前&…...

hcip实验2
根据地址分配完成基础配置 先配置r1,r2,r3的ospf以及与isp通讯: 配置缺省路由: 完成nat配置: 完成r5,r6,r7,r8,r15的mgre以及整个网络的ospf配置 mgre: area 2 和3之间用多进程双向重发布技术完成: area4和5之间用虚…...