ort 2.2.99 cannot be built on PPC Mac OS X 10.4.11 (Tiger) because of SYS_lstat64 undeclared in lstat64()

[ballapete (Peter "Pete" Dyballa)]

cc -g -O2 -std=c99 -Wextra -Wall   -fPIC -arch i386 -arch ppc -DHAVE_CONFIG_H -I.. -I.  -I"/usr/include"  -c -o stat.o stat.c
stat.c: In function ‘stat64’:
stat.c:79: error: ‘SYS_stat64’ undeclared (first use in this function)
stat.c:79: error: (Each undeclared identifier is reported only once
stat.c:79: error: for each function it appears in.)
stat.c: In function ‘lstat64’:
stat.c:121: error: ‘SYS_lstat64’ undeclared (first use in this function)
stat.c: In function ‘stat64’:
stat.c:79: error: ‘SYS_stat64’ undeclared (first use in this function)
stat.c:79: error: (Each undeclared identifier is reported only once
stat.c:79: error: for each function it appears in.)
stat.c: In function ‘lstat64’:
stat.c:121: error: ‘SYS_lstat64’ undeclared (first use in this function)
lipo: can't figure out the architecture type of: /var/tmp//ccBst4UU.out
make[3]: *** [stat.o] Error 1
make[2]: *** [all] Error 1
make[1]: *** [all] Error 1
make: *** [all] Error 2

This happen on a G4 (PPC 7447A) with 32-bit hardware. (And software.)

[neverpanic (Clemens Lang)]

Can you check whether r113871, r113872 and r113873 fix the build on Tiger for you?

[ballapete (Peter "Pete" Dyballa)]

Replying to cal@…:

Can you check whether r113871, r113872 and r113873 fix the build on Tiger for you?

No, a previously reported bug is back:

make[2]: Nothing to be done for `all'.
===> making all in src/pextlib1.0
cc -c -DUSE_TCL_STUBS -DTCL_NO_DEPRECATED -g -O2 -std=c99 -Wextra -Wall -pedantic     -DHAVE_CONFIG_H -I.. -I.  -I"/usr/include" -fno-common tracelib.c -o tracelib.o
tracelib.c: In function ‘TracelibRunCmd’:
tracelib.c:585: error: ‘EV_RECEIPT’ undeclared (first use in this function)
tracelib.c:585: error: (Each undeclared identifier is reported only once
tracelib.c:585: error: for each function it appears in.)
make[2]: *** [tracelib.o] Error 1
make[1]: *** [all] Error 1
make: *** [all] Error 1

Configure correctly saw that DTrace is not available, but then this slipped in…

[neverpanic (Clemens Lang)]

Can you attach config.log?

[ballapete (Peter "Pete" Dyballa)]

config.log

[ballapete (Peter "Pete" Dyballa)]

Replying to cal@…:

Can you attach config.log?

Done!

[neverpanic (Clemens Lang)]

It seems the check works correctly:

configure:9815: checking whether trace mode is supported on this platform
configure:9850: result: EV_RECEIPT not available, no

HAVE_TRACEMODE_SUPPORT is defined to 0 on your system – it shouldn't be defined at all, though. My mistake.

[neverpanic (Clemens Lang)]

r113877 and r113878.

[ballapete (Peter "Pete" Dyballa)]

Replying to cal@…:

r113877 and r113878.

Success!

Only some tests from 'make test' fail:

===> making test in src/macports1.0
/usr/bin/tclsh ./tests/test.tcl -nocolor
Warning: test
Total:39 Passed:34 Failed:05 Skipped:00  macports.test
--->  Fetching distfiles for fondu
--->  Attempting to fetch fondu_src-060102.tgz from http://fondu.sourceforge.net/
--->  Verifying checksums for fondu
--->  Extracting fondu
--->  Configuring fondu
--->  Building fondu
--->  Staging fondu into destroot
--->  Uninstalling fondu @3.0_1

==== set_developer_dir Set developer dir unit test. Tests only for correct xcode-select dir. FAILED
==== Contents of test case:

    unset macports::developer_dir

    if {[macports::set_developer_dir a b c] != ""} {
       return "FAIL: cannot set dev dir"
    }
    if {![info exists macports::developer_dir]} {
       return "FAIL: developer_dir var no set"
    }
    return "Set developer dir successful."

---- Result was:
FAIL: cannot set dev dir
---- Result should have been (exact matching):
Set developer dir successful.
==== set_developer_dir FAILED


==== copy_xcode_plist Copy xcode plist unit test. FAILED
==== Contents of test case:

    set target $pwd/target

    if {[macports::copy_xcode_plist $target] != ""} {
       return "FAIL: cannot copy xcode plist"
    }
    if {![file exists $target/Library/Preferences/com.apple.dt.Xcode.plist]} {
       return "FAIL: missing plist file"
    }
    return "Copy xcode plist successful."

---- Result was:
FAIL: missing plist file
---- Result should have been (exact matching):
Copy xcode plist successful.
==== copy_xcode_plist FAILED


==== worker_init Worker init unit test. FAILED
==== Contents of test case:

    macports::worker_init $name $portpath $porturl $portbuildpath $options $variations
    if {$name != "interp0"} {
       return "FAIL: wrong workername"
    }
    if {[$name eval source Portfile] != "yes"} {
       return "FAIL: cannot load Portfile"
    }
    if {[$name eval findBinary ls] != "/bin/ls"} {
       return "FAIL: alias not created"
    }
    if {[$name eval return \$os_arch] != "i386"} {
       return "FAIL: var not set"
    }
    return "Worker init successful."

---- Result was:
FAIL: var not set
---- Result should have been (exact matching):
Worker init successful.
==== worker_init FAILED


==== fetch_port Fetch port unit test. FAILED
==== Contents of test case:

    set url http://packages.macports.org/fondu/fondu-060102_1.darwin_10.x86_64.tbz2
    set res [macports::fetch_port $url]
    if {$res != "${pwd}/portdbpath/portdirs/fondu-060102_1"} {
	return "FAIL: cannot fetch port"
    }
    return "Fetch port successful."

---- Result was:
FAIL: cannot fetch port
---- Result should have been (exact matching):
Fetch port successful.
==== fetch_port FAILED


==== revupgrade_handle_special_paths Revupgrade handle special paths unit test. FAILED
==== Contents of test case:

    set res [macports::revupgrade_handle_special_paths fname test_path]
    if {$res != "test_path"} {
       return "FAIL: wrong path"
    }
    set res [macports::revupgrade_handle_special_paths fname @loader_path/test_load]
    if {$res != "./test_load"} {
       return "FAIL: wrong load path"
    }
    return "Revupgrade handle special path successful."

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: bad index "0+11": must be integer or end?-integer?
    while executing
"string replace $corrected_path $loaderpath_idx ${loaderpath_idx}+11 [file dirname $fname]"
    (procedure "macports::revupgrade_handle_special_paths" line 6)
    invoked from within
"macports::revupgrade_handle_special_paths fname @loader_path/test_load"
    invoked from within
"set res [macports::revupgrade_handle_special_paths fname @loader_path/test_load]"
    ("uplevel" body line 6)
    invoked from within
"uplevel 1 $script"
---- errorCode: NONE
==== revupgrade_handle_special_paths FAILED


Total:14 Passed:14 Failed:00 Skipped:00  macports_dlist.test
Total:08 Passed:07 Failed:01 Skipped:00  macports_util.test

==== method_wrap Method wrap unit test. FAILED
==== Contents of test case:

    macports_util::method_wrap test_proc
    if {[test_proc 1] != 7} {
       return "FAIL: no wrap around method"
    }
    if {[test_proc arg] != "arg"} {
       return "FAIL: no wrap around method"
    }
    return "Method wrap successful."

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: bad option "1": must be -code, -errorcode, or -errorinfo
    while executing
"test_proc 1"
    ("uplevel" body line 1)
---- errorCode: NONE
==== method_wrap FAILED


make[2]: *** [test] Error 1
make[1]: *** [test] Error 1
make: *** [test] Error 1

Compilation exited abnormally with code 2 at Mon Nov 25 01:02:46

[neverpanic (Clemens Lang)]

I think you can safely ignore those for now, since the tests are rather new. I'll try to have a look.

[cooljeanius (Eric Gallager)]

Cc Me!

[neverpanic (Clemens Lang)]

Please ignore the test failures for now. I don't think Tiger is relevant enough that I'll ever fix those.