Hi,
let me say it again: never return a null.
One more thing: my implementation uses a trivial mechanism to make sure
that there is only one instance of the service (I would have to store
the state on disk otherwise). If your implementation is stateless, you
might have many instances, but in case you want to save the ignored
rules or save memory for instantiated objects, it might be good idea to
have a single thread only.
Is your source available in CVS? I might look at it.
Regards
Marcin
William Colen pisze:
Thank you Thomas,
I can see the call to getLocales method in our extension logs. Maybe
for some threads it is working, for other threads not?
I compared our implementation with Marcin's implementation. The only
diference is that he passed a empty string as the 3rd parameter, we
passed null.
Maybe that would solve this issue, but I will only be able to change
that and recompile at night (UTC/GMT -2 hours)
Thank you!
William
On Fri, Jan 9, 2009 at 8:53 AM, Thomas Lange - Sun Germany - ham02 -
Hamburg <[email protected]> wrote:
Hello William and Carlos,
I was able to abtain the implementation name
"br.usp.pcs.lta.cogroo.ooointegration.Main" by debugging linguistic
(namely LngSvcMgr::GetAvailableGrammarSvcs_Imp) the grammar checker
itself can be instantiated. It is even possible to get the
XSupportedLocales interface from the component but the call to
getLocales() does not work properly.
In the debugger I can not step properly over the line
uno::Sequence<lang::Locale> aLocaleSequence(xSuppLoc->getLocales());
and get a result from that call in LngSvcMgr::GetAvailableGrammarSvcs_Imp.
Thus I think you should start by checking that implementation.
While stepping over that call I'll get the following:
First-chance exception at 0x77e4bee7 in soffice.bin: Microsoft C++
exception: jni_uno::BridgeRuntimeError at memory location 0x014bd974
Of course there might be other problems later on, but since I could not
properly get beyond that point looking into problems that are beyond
that point will probably be in vain as long that call does not work.
And just in case that I have not mentioned this elsewhere so far:
It is absolutely necessary that the component is thread-safe since it
will be called from different threads!
Regards,
Thomas
Hi William,
William Colen wrote:
Thomas,
I uploaded the extension to
https://sourceforge.net/project/showfiles.php?group_id=182075&package_id=305572
The file is
CoGrOO-AddOn-3.0.2-SNAPSHOT-bin.oxt<http://downloads.sourceforge.net/cogroo/CoGrOO-AddOn-3.0.2-SNAPSHOT-bin.oxt?use_mirror=>
Thank you!
Marcin,Thank you! The tutorial helped a lot! I could debug it.
But I still not able to find the crash problem.
Following you can see the logs generated by CoGrOO extension.
My steps to cause the crash was:
1. Created a new document.
2. Changed it to pt_BR
3. Entered a sentence with a grammar error: Fomos levados à crer.
4. Waited a few seconds, no automatic grammar checking (all log entries I
have was generated before I press F7)
5. Pressed F7 -> OOo crashed
The log is clear and easy to read. Something interesting is that the
method doProofreading was never called: no traces from that method and also
I inserted a breakpoint at the method's first line, and Eclipse never
stopped there.
Can you also drop me the implementation name of the service?
Since already the Tools/Options/Writing-Aids dialog does not get
displayed anymore I suspect I may have to do with the registration of
the service. Thus I like to instantiate it manually via API...
Also you need not start the checking with F7 to crash the Office it
already dies just by trying to opening the Format/Character dialog. :-(
Regards,
Thomas
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
