This looks like an invalid pointer issue, the _4294967295 should be returned from a previous ole call. You should be able to run J32 on your 64-bit windows 8.1, can you also try that? Testing with jconsole only should be sufficient.
Чт, 06 ноя 2014, Mike Day написал(а): > OK - progress in some ways. As Bill is running LibreOffice, I > downloaded it in place of Open Office, tweaking the default > program settings to force wdooo calls to invoke its binaries. > > My application still crashes with the new JQT and Jconsole, > and so does oocalc.ijs. > > HOWEVER, I've changed oleinvoke from an adverb to a > verb, together with the verbs that call it. This is so that > debug can trace through oleinvoke (is there a way to trace > execution through a verb derived on the fly from an adverb?) > > So now, I can report that test'' fails at the following line in oleinvoke: > (NB. I'm breaking it into three parts, so it should appear as three lines > here) > > if. > S_OK~: hr=. disp idInvoke id ; GUID_NULL ; 0 ; m ; (<dispparams) ; (<temp) ; > 0 ; 0 > do. 13!:8[3 [ oleerrno=: hr > > The rh argument for idInvoke is > id ; GUID_NULL ; 0 ; m ;(<dispparams) ; (<temp) ; 0 ; 0 > > +-----------+----------------+-+-+--------------+--------------+-+-+ > > |_4294967295| |0|1|+------------+|+------------+|0|0| > > | | | | ||867122743072|||867151675840|| | | > > | | | | |+------------+|+------------+| | | > > +-----------+----------------+-+-+--------------+--------------+-+-+ > > (though presumably handles are arbitrary) > > Opening the two boxed items makes this easier to read: > id ; GUID_NULL ; 0 ; m ;dispparams ; temp ; 0 ; 0 > > +-----------+----------------+-+-+------------+------------+-+-+ > > |_4294967295| |0|1|867122743072|867151675840|0|0| > > +-----------+----------------+-+-+------------+------------+-+-+ > > (Apologies if these boxes are poorly aligned!) > > > Copying the following from the debug gui and running explicitly > > in the J terminal session > disp idInvoke id ; GUID_NULL ; 0 ; m ; (<dispparams) ; (<temp) ; 0 ; 0 > leads to the crash. > > > idInvoke is a system call: > '1 6 > i x x *c x s *x *x x x'&(15!:0)@; > > > That's all folks! > > > Thanks, Mike > > > > On 06/11/2014 16:09, Mike Day wrote: > >But how to see them if J crashes first? > > > >I could see that S_OK was zero in olecreate, > >but debug then only got as far as the first line > >oleinvoke with no chance of checking the code. > > > >Bill was keen to see reports on wdooo behaviour > >with a variety of Windows o/s and numbers of > >bits. > > > >Mike > > > >On 06/11/2014 13:37, bill lam wrote: > >>When calls in wdooo fail, they should return non-zero error code,that > >>might > >>give hints of why it failed. > >>On Nov 6, 2014 5:05 PM, "Mike Day" <[email protected]> wrote: > >> > >>>Dear Forum > >>> > >>>I've got a problem with J crashing with wdooo, which I've > >>>been discussing privately with Bill Lam. Privately because > >>>my example involved private data. > >>> > >>>We seem to have discovered where the trouble lies or at > >>>least where it starts, and in a public area, so it can be > >>>shared. > >>> > >>>I tried the verb "test" in ~addons/tables/wdooo/oocalc.ijs. > >>>Using debug, I see that it fails at the line > >>>olemethod__p base ; 'createInstance' ; 'com.sun.star.frame.Desktop' > >>>(Actually, olemethod attempts to call oleinvoke; the error appears > >>>to occur with '' (m oleinvoke) y ) > >>> > >>> > >>>Executing this line throws an exception under Windows 8.1 > >>>with the message > >>> jqt.exe has stopped working etc > >>> > >>>While I'm running Apache Open Office 4.1.1 in Windows 8.1 on a > >>>64 bit machine, Bill has libreoffice 4.3.3 in windows xp in > >>>32 bits, where oocalc's test works ok, so we're at an impasse. > >>> > >>>Is anyone else who's got an Open Office + 64-bit Windows 8.1 > >>>setup willing to try the oocalc test? > >>> > >>>I recall trying out libreoffice two or three years ago and then > >>>finding Open Office to be more stable - but perhaps things have > >>>improved. > >>> > >>>Thanks in advance, > >>> > >>>Mike > > > --- > This email is free from viruses and malware because avast! Antivirus > protection is active. > http://www.avast.com > > > > ----- > No virus found in this message. > Checked by AVG - www.avg.com > Version: 2014.0.4765 / Virus Database: 4189/8522 - Release Date: 11/06/14 > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm -- regards, ==================================================== GPG key 1024D/4434BAB3 2008-08-24 gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3 gpg --keyserver subkeys.pgp.net --armor --export 4434BAB3 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
