harfbuzz @2.3.0 build and test errors on some legacy systems

[kencu (Ken)]

In the process of working with harfbuzz recently, I came across various build and test errors that might be of interest.

building with gcc-4.2, so far tested on 10.4, 10.5 (intel or PPC) errors due to:

/usr/bin/make  check-TESTS
PASS: check-c-linkage-decls.sh
PASS: check-externs.sh
PASS: check-header-guards.sh
PASS: check-includes.sh
SKIP: check-static-inits.sh
FAIL: check-symbols.sh
PASS: check-libstdc++.sh
make[6]: Nothing to be done for `all'.
PASS: test-ot-tag
PASS: test-unicode-ranges
============================================================================
Testsuite summary for HarfBuzz 2.3.0
============================================================================
# TOTAL: 9
# PASS:  7
# SKIP:  1
# XFAIL: 0
# FAIL:  1
# XPASS: 0
# ERROR: 0

building with gcc6 on 10.4 and 10.5 PPC succeeds and passes all tests

building with gcc6 on 10.5 Intel fails:

In file included from /System/Library/Frameworks/Security.framework/Headers/Security.h:57:0,
                 from /System/Library/Frameworks/CoreServices.framework/Frameworks/LaunchServices.framework/Headers/LSSharedFileList.h:32,
                 from /System/Library/Frameworks/CoreServices.framework/Frameworks/LaunchServices.framework/Headers/LaunchServices.h:37,
                 from /System/Library/Frameworks/CoreServices.framework/Headers/CoreServices.h:41,
                 from /System/Library/Frameworks/ApplicationServices.framework/Headers/ApplicationServices.h:20,
                 from hb-coretext.h:37,
                 from hb-coretext.cc:32:
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:102:46: error: shift expression '(1853123693 << 8)' overflows [-fpermissive]
     kSecAuthenticationTypeNTLM             = AUTH_TYPE_FIX_ ('ntlm'),
                                              ^
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:102:46: error: enumerator value for 'kSecAuthenticationTypeNTLM' is not an integer constant
     kSecAuthenticationTypeNTLM             = AUTH_TYPE_FIX_ ('ntlm'),
                                              ^
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:103:46: error: shift expression '(1836281441 << 8)' overflows [-fpermissive]
     kSecAuthenticationTypeMSN              = AUTH_TYPE_FIX_ ('msna'),
                                              ^
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:103:46: error: enumerator value for 'kSecAuthenticationTypeMSN' is not an integer constant
     kSecAuthenticationTypeMSN              = AUTH_TYPE_FIX_ ('msna'),
                                              ^
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:104:46: error: shift expression '(1685086561 << 8)' overflows [-fpermissive]
     kSecAuthenticationTypeDPA              = AUTH_TYPE_FIX_ ('dpaa'),
                                              ^
/System/Library/Frameworks/Security.framework/Headers/SecKeychain.h:104:46: error: enumerator value for 'kSecAuthenticationTypeDPA' is not an integer constant
     kSecAuthenticationTypeDPA              = AUTH_TYPE_FIX_ ('dpaa'),

but building with gcc6 on 10.5 Intel (i386) succeeds with the following flag added:

configure.cxxflags-append -fpermissive

and passes all tests

building with clang-3.4 or clang-5.0 on 10.5 Intel (i386) against libstdc++ fails tests:

FAIL: tests/kern-format2.tests

same happens with PortGroup cxx11 1.1 building against macports-libstdc++

building with clang-5.0 against libc++ on 10.6.8, and with XCode clang on 10.13 and 10.14 passes all tests. 10.7 to 10.12 not yet tested.

Where is this going? It looks like harfbuzz is a bit touchy about exactly how it's built, at least from the perspective of the test suite.

[kencu (Ken)]

Data point:

Building harfbuzz 2.5.3 with clang-5.0 against the macports-libstdc++ headers using the PG cxx11 1.1 on 10.7.5 passes all tests. This is the current default for 10.6 to 10.8.

(64bit only tested.)

[kencu (Ken)]

there's nothing to fix here.