On Tuesday November 24 2015 20:00:27 Clemens Lang wrote:
> ----- On 24 Nov, 2015, at 10:45, René J.V. Bertin [email protected] wrote:
>
> > Turns out it links to libboost_program_options, but rev-upgrade didn't
> > catch any
> > issues.
> >
> > Should it have had?
>
> Depends. Did the compatibility version of libboost_program_options increase to
> indicate that the library is no longer compatible to its previous version?
Ah, no, both identify as
/opt/local/lib/libboost_program_options-mt.dylib (compatibility version 0.0.0,
current version 0.0.0)
>From what I can tell, I had built akonadi against boost 1.56.0 (port rev. 2).
> Also, how did it print a backtrace? If the library could not be loaded you
> should
>From the looks of it backtrace() and backtrace_symbols() were called somewhere
>(without subsequent demangling step).
It was clearly not an error from dyld.
> A third possibility would be that boost was built against a different C++
> runtime
> library than akonadi.
Unlikely on 10.9 ...
> The backtrace would tell us more.
Fortunately I still had 2 in the terminal scrollback buffer:
#> akonadictl start
"[
0: 0 akonadictl 0x000000010c1239cc _Z11akBacktracev
+ 76
1: 1 akonadictl 0x000000010c123d0c
_ZL19defaultCrashHandleri + 92
2: 2 libsystem_platform.dylib 0x00007fff89c3f5aa _sigtramp + 26
3: 3 QtDBus 0x000000010c582b5a
_ZZN22QDBusConnectionPrivate14registerObjectEPKNS_14ObjectTreeNodeEE7counter +
16922
4: 4 libboost_program_options-mt.dylib 0x000000010c6217a4
_ZN5boost15program_options5storeERKNS0_20basic_parsed_optionsIcEERNS0_13variables_mapEb
+ 1956
5: 5 akonadictl 0x000000010c11b2c2 main + 3346
6: 6 libdyld.dylib 0x00007fff90ce65fd start + 1
7: 7 ??? 0x0000000000000002 0x0 + 2
]
"
Exit 255
#> akonadictl restart
"[
0: 0 akonadictl 0x000000010f7aa9cc _Z11akBacktracev
+ 76
1: 1 akonadictl 0x000000010f7aad0c
_ZL19defaultCrashHandleri + 92
2: 2 libsystem_platform.dylib 0x00007fff89c3f5aa _sigtramp + 26
3: 3 QtDBus 0x000000010fc09b5a
_ZZN22QDBusConnectionPrivate14registerObjectEPKNS_14ObjectTreeNodeEE7counter +
16922
4: 4 libboost_program_options-mt.dylib 0x000000010fca37a4
_ZN5boost15program_options5storeERKNS0_20basic_parsed_optionsIcEERNS0_13variables_mapEb
+ 1956
5: 5 akonadictl 0x000000010f7a22c2 main + 3346
6: 6 libdyld.dylib 0x00007fff90ce65fd start + 1
7: 7 ??? 0x0000000000000002 0x0 + 2
]
"
Which translates into
0: 0 akonadictl 0x000000010f7aa9cc akBacktrace() + 76
1: 1 akonadictl 0x000000010f7aad0c
defaultCrashHandler(int) + 92
2: 2 libsystem_platform.dylib 0x00007fff89c3f5aa _sigtramp + 26
3: 3 QtDBus 0x000000010fc09b5a
QDBusConnectionPrivate:: \
registerObject(QDBusConnectionPrivate::ObjectTreeNode const*)::counter + 16922
4: 4 libboost_program_options-mt.dylib 0x000000010fca37a4
boost::program_options:: \
store(boost::program_options::basic_parsed_options<char> const&, \
boost::program_options::variables_map&, bool)
(and so akonadi provides its own default crashhandler that prints a backtrace.
Neat.)
Also looks like I was lucky that rebuilding against the new boost version was
enough, but then again Qt 4 hasn't changed at all...
R.
_______________________________________________
macports-dev mailing list
[email protected]
https://lists.macosforge.org/mailman/listinfo/macports-dev
