[interfacekit] Re: Broken BMessage
- From: Erik Jaesler <erik@xxxxxxxxxxxxxx>
- To: interfacekit@xxxxxxxxxxxxx
- Date: Fri, 08 Oct 2004 09:58:08 -0700
Following Ingo's very clear instructions, I was able to confirm that any
BMessage-related issues with BAppFileInfo appear to be fixed. There are
still unit test failures, but these have to do with MIME type
installation failing. The problem appears to be in Database::Install()
(in /src/kits/storage/mime/Database.cpp); if
/boot/home/config/settings/obos_mime/application (I think that's what
it's called) doesn't exist, the function simply errs out, rather than
creating the directory, which seems to me the appropriate thing to do.
Obviously, Storage Kit is not my baliwick, so I haven't attempted to
make a fix/*really* messed things up. ;)
I'll be trying to find time to diagnose and fix the remaining issues
next week.
Thanks,
e
Erik Jaesler wrote:
I finally found the time to do some work. Fixed the SendMessageTest1()
issue (BMessage::fWasDelivered wasn't getting initialized when
unflattening); that now passes. Unfortunately, the OBOS version of the
unit tests for storage kit don't seem to get loaded, though everything
looks correct: the lib is in the right place and is exporting the right
entry point. If anybody has any idea why that might be, I'd love some
suggestions; otherwise I'll have to do a debug build of the testing
framework and step through. =P
Couldn't get anywhere with the BApplication tests, on account of
app_server dying on a strcmp() in CursorManager::RemoveAppCursors()
during QuitTest2(). Anyway, still pokin' away here when time presents
itself. =)
e
The BNodeInfo test still crashes the registrar:
loading symbols
segment violation occurred
memcpy_pIII:
memcpy_pIII:
+0047 ec0912e7: * 068a movb (%esi), %al
main_mime:sc
frame retaddr
fd081fac ecfb43c4 BPrivate::BDataBuffer::BDataReference::Create(void
*,
unsigned long, BPrivate::BDataBuffer::BDataReference *&, bool) +
00000038
fd081fd8 ecfb42c6 BPrivate::BDataBuffer::BDataBuffer(void *,
unsigned
long, bool) + 0000002a
fd081ff8 ecfbe767 BMessage::AddData(char const *, unsigned long,
void
const *, long, bool, long) + 00000313
fd082034 ecfbcc16 BMessage::Unflatten(BDataIO *) + 0000060a
fd08219c ecfbc5b4 BMessage::Unflatten(char const *) + 00000030
fd0821f0 ecfba1d6 BLooper::ConvertToMessage(void *, long) + 00000066
fd082214 ecfba152 BLooper::ReadMessageFromPort(long long) + 00000046
fd082248 ecfb99f9 BLooper::MessageFromPort(long long) + 00000021
fd082260 ecfba284 BLooper::task_looper(void) + 00000048
fd08228c ecfba05a BLooper::_task0_(void *) + 00000036
fd0822a0 ec0851ed thread_start + 00000039
main_mime:
The BAppFileInfo unit test doesn't crash anymore in the "Supported
Types
Test", but the test case fails and the following test case ("Icon
Test")
hangs. It seems like the "main_mime" thread (a BLooper thread) gets
terminated somehow and the test app is waiting for a reply on a
temporary
message port for good. The same holds for the BMimeType unit test.
The BApplication unit test (the remote application, to be precise)
crashs:
loading symbols
segment violation occurred
BMessage::Unflatten(BDataIO *):
Unflatten__8BMessageP7BDataIO:
+05d3 ecfbcbdf: * 068b movl (%esi), %eax
AppRunTestApp1:sc
frame retaddr
fd0015d8 ecfbc5b4 BMessage::Unflatten(char const *) + 00000030
fd00162c ecfba1d6 BLooper::ConvertToMessage(void *, long) + 00000066
fd001650 ecfba152 BLooper::ReadMessageFromPort(long long) + 00000046
fd001684 ecfb99f9 BLooper::MessageFromPort(long long) + 00000021
fd00169c ecfba284 BLooper::task_looper(void) + 00000048
fd0016c8 ecfb2780 BApplication::run_task(void) + 00000028
fd0016d8 ecfb188e BApplication::Run(void) + 00000056
fd0016f0 80002809 CommonTestApp::Run(void) + 0000001d
fd001708 800023ae main + 0000008e
fd001844 80002205 _start + 00000061
AppRunTestApp1:
Furthermore the BMessenger::SendMessageTest1 test case fails.
CU, Ingo
Necessity is the plea for every infringement of human freedom. It is the
argument of tyrants; it is the creed of slaves.
-William Pitt, British prime-minister (1759-1806)
- Follow-Ups:
- [interfacekit] Re: Broken BMessage
- From: Ingo Weinhold
- References:
- [interfacekit] Re: Broken BMessage
- From: Erik Jaesler
Other related posts:
- » [interfacekit] Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
- » [interfacekit] Re: Broken BMessage
Couldn't get anywhere with the BApplication tests, on account of app_server dying on a strcmp() in CursorManager::RemoveAppCursors() during QuitTest2(). Anyway, still pokin' away here when time presents itself. =)
e
The BNodeInfo test still crashes the registrar:
loading symbols
segment violation occurred
memcpy_pIII:
memcpy_pIII:
+0047 ec0912e7: * 068a movb (%esi), %al
main_mime:sc
frame retaddr
fd081fac ecfb43c4 BPrivate::BDataBuffer::BDataReference::Create(void
*,
unsigned long, BPrivate::BDataBuffer::BDataReference *&, bool) +
00000038
fd081fd8 ecfb42c6 BPrivate::BDataBuffer::BDataBuffer(void *,
unsigned
long, bool) + 0000002a
fd081ff8 ecfbe767 BMessage::AddData(char const *, unsigned long,
void
const *, long, bool, long) + 00000313 fd082034 ecfbcc16 BMessage::Unflatten(BDataIO *) + 0000060a fd08219c ecfbc5b4 BMessage::Unflatten(char const *) + 00000030 fd0821f0 ecfba1d6 BLooper::ConvertToMessage(void *, long) + 00000066 fd082214 ecfba152 BLooper::ReadMessageFromPort(long long) + 00000046 fd082248 ecfb99f9 BLooper::MessageFromPort(long long) + 00000021 fd082260 ecfba284 BLooper::task_looper(void) + 00000048 fd08228c ecfba05a BLooper::_task0_(void *) + 00000036 fd0822a0 ec0851ed thread_start + 00000039 main_mime:
The BAppFileInfo unit test doesn't crash anymore in the "Supported
Types
Test", but the test case fails and the following test case ("Icon
Test")
hangs. It seems like the "main_mime" thread (a BLooper thread) gets terminated somehow and the test app is waiting for a reply on a
temporary
message port for good. The same holds for the BMimeType unit test.
The BApplication unit test (the remote application, to be precise)
crashs:
loading symbols segment violation occurred BMessage::Unflatten(BDataIO *): Unflatten__8BMessageP7BDataIO: +05d3 ecfbcbdf: * 068b movl (%esi), %eax AppRunTestApp1:sc frame retaddr fd0015d8 ecfbc5b4 BMessage::Unflatten(char const *) + 00000030 fd00162c ecfba1d6 BLooper::ConvertToMessage(void *, long) + 00000066 fd001650 ecfba152 BLooper::ReadMessageFromPort(long long) + 00000046 fd001684 ecfb99f9 BLooper::MessageFromPort(long long) + 00000021 fd00169c ecfba284 BLooper::task_looper(void) + 00000048 fd0016c8 ecfb2780 BApplication::run_task(void) + 00000028 fd0016d8 ecfb188e BApplication::Run(void) + 00000056 fd0016f0 80002809 CommonTestApp::Run(void) + 0000001d fd001708 800023ae main + 0000008e fd001844 80002205 _start + 00000061 AppRunTestApp1:
Furthermore the BMessenger::SendMessageTest1 test case fails.
CU, Ingo
Necessity is the plea for every infringement of human freedom. It is the argument of tyrants; it is the creed of slaves.
-William Pitt, British prime-minister (1759-1806)
- [interfacekit] Re: Broken BMessage
- From: Ingo Weinhold
- [interfacekit] Re: Broken BMessage
- From: Erik Jaesler