Date: Sat, 10 Sep 2011 07:50:08 GMT From: "Hartmann, O." <ohartman@zedat.fu-berlin.de> To: office@FreeBSD.org Subject: Re: ports/160481: devel/icu: Errors in total: 1. TestCompilerRTTI UObjectTest utility CLANG build failure Message-ID: <201109100750.p8A7o88s081152@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR ports/160481; it has been noted by GNATS.
From: "Hartmann, O." <ohartman@zedat.fu-berlin.de>
To: Sunpoet Po-Chuan Hsieh <sunpoet@FreeBSD.org>
Cc: bug-followup@FreeBSD.org, office@FreeBSD.org
Subject: Re: ports/160481: devel/icu: Errors in total: 1. TestCompilerRTTI
UObjectTest utility CLANG build failure
Date: Sat, 10 Sep 2011 09:47:34 +0200
On 09/09/11 22:16, Sunpoet Po-Chuan Hsieh wrote:
> On Mon, Sep 05, 2011 at 09:18:25PM +0200, Hartmann, O. wrote:
>> On 09/05/11 18:16, Sunpoet Po-Chuan Hsieh wrote:
>>> Hi,
>>>
>>> The failure is not limited to clang. It will also occur if you use gcc. :)
>>>
>>> excerpt from ${WRKSRC}/test/intltest/uobjtest.cpp:
>>>
>>> NumberFormat *nf = NumberFormat::createInstance("de", errorCode);
>>>
>>> ...
>>>
>>> if (&typeid(*nf) == NULL || typeid(*nf) == typeid(UObject) || typeid(*nf) == typeid(Format) ||
>>> typeid(*nf) != typeid(DecimalFormat) || typeid(*nf) == typeid(ChoiceFormat) ||
>>> typeid(*nf) == typeid(emptySet)
>>> ) {
>>> errln("typeid(NumberFormat) failed");
>>> }
>>>
>>> The failure was raised by typeid(*nf) != typeid(DecimalFormat).
>>> However, according to [1], I think it should be ==, not !=.
>>> That means you could safely ignore this. :)
>>>
>>> [1] http://icu-project.org/apiref/icu4c/classNumberFormat.html
>>>
>>> Regards,
>> Hello.
>> Using the legacy system's gcc 4.2, the port devel/icu compiles and
>> installs without error.
>> Using clang, the port is not installed properly.
>>
>> Using a portmaster -f xorg, which triggers the build of all ports
>> neccessary for xorg ports
>> matched by the pattern "xorg" results in a stop of the build
>> process. I can not test further
>> the proper build of the X11 subsystem without excluding explicitely
>> devel/icu (amongst others).
>>
>> The error occured does not seem to be "obvious" to me, so reporting
>> this issue by a PR seems the proper
>> way.
>>
>> Can this be fixed so that clang compiles the port and portmaster
>> installes the port properly?
>>
>> Regards,
>> Oliver
> Please try the patch at http://people.freebsd.org/~sunpoet/icu.patch
> Thanks.
>
> Regards,
Hello.
Doesn't work. After application of the suggested patch I receive this:
} OK: rbnfp
csdet {
ConstructionTest {
} OK: ConstructionTest
UTF8Test {
} OK: UTF8Test
UTF16Test {
} OK: UTF16Test
C1BytesTest {
} OK: C1BytesTest
InputFilterTest {
} OK: InputFilterTest
DetectionTest {
} OK: DetectionTest (68ms)
IBM424Test {
} OK: IBM424Test
IBM420Test {
} OK: IBM420Test
Ticket6394Test {
} OK: Ticket6394Test
} OK: csdet (70ms)
spoof {
TestSpoofAPI {
} OK: TestSpoofAPI
TestSkeleton {
} OK: TestSkeleton
TestAreConfusable {
} OK: TestAreConfusable
TestInvisible {
} OK: TestInvisible
testConfData {
} OK: testConfData (76ms)
testBug8654 {
} OK: testBug8654
} OK: spoof (76ms)
bidi {
TestBidiTest {
} OK: TestBidiTest (250ms)
} OK: bidi (250ms)
--------------------------------------
Errors in total: 1.
TestCompilerRTTI
UObjectTest
utility
--------------------------------------
Elapsed Time: 00:01:38.808
*** Error code 1
Stop in /usr/ports/devel/icu.
*** Error code 1
Stop in /usr/ports/devel/icu.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201109100750.p8A7o88s081152>