Steps to reproduce: (1) Open a new Writer document. (2) Open the Basic IDE and navigate to the source code of a module of the current file: (2.1) Take menu options Tools > Macros > "Organize macros" > "LibreOfficeDev Basic...". Program presents dialog "LibreOfficeDev Basic Macros". (2.2) In the left pane, "Macro from", click "Untitled 1" and then click <New>. LibreOffice presents dialog "New Module" with Name defaulted to "Module1". (2.3) Click <OK>. Program presents window "Untitled 1.Standard" (3) Into the source pane paste: sub crash dim price as double price = 8.6 print price end sub 'crash (4) Place the caret on the line "sub crash" and type <F5> to run the macro. Result expected : message box saying "8.6" Result observed : SIGABRT. These observations are on master commit 480c7c2, fetched 2013-12-127 02:33 UTC, configured as: --enable-option-checking=fatal --enable-dbgutil --enable-crashdump --without-system-postgresql --without-myspell-dicts --with-extra-buildid --without-doxygen --with-external-tar=/home/terry/lo_hacking/git/src built and running on debian-wheezy 64-bit. Note for comparison that version 4.2.0.1 displays the message box that I expected. I cannot tell whether the difference is a recent commit to master or my debug build. I shall soon attach a typescript with a backtrace obtained from the above steps. Note that the throw comes from framework::XMLBasedAcceleratorConfiguration::getCommandByKeyEvent, framework/source/accelerators/acceleratorconfiguration.cxx:126. Just as an aside, and to head off those who would ask me to attach a document demonstrating the crash ... I tried to prepare a Writer document with a button invoking the macro 'crash'. It only crashed sometimes, and when it did crash it crashed after different numbers of throws executed and with different backtraces. The STR above, by contrast have produced essentially identical results three times in a row; I lack the courage to try for a fourth time <grin />.
Created attachment 91395 [details] typescript with backtrace with symbols Some line numbers within the typescript: 24: start of messages from Libreoffice 114: backtrace from the last `throw` 154: backtrace from catch 196: backtrace frm SIGABRT
I observe the error message (BASIC-runtime-error [...] com.sun.star.container.NoSuchElementException, message: .) with version 4.2.2.1 when saving a writer document that has BASIC macros. Error message does not occur with version 4.1.5. Error message does occur with version 4.3.0.0-alpha0+ (master, last pull yesterday). I cannot reproduce the error with the steps in comment#1, neither with version 4.2.2 nor with master. Both give 8.6 when running the macro.
(In reply to comment #2) > I observe the error message (BASIC-runtime-error [...] > com.sun.star.container.NoSuchElementException, message: .) > with version 4.2.2.1 when saving a writer document that has BASIC macros. > > Error message does not occur with version 4.1.5. > Error message does occur with version 4.3.0.0-alpha0+ (master, last pull > yesterday). > > I cannot reproduce the error with the steps in comment#1, neither with > version 4.2.2 nor with master. Both give 8.6 when running the macro. Our problem has been 'fixed' by changing the macro, which was looking for a table with a non-existing name. As the problem did not occur with version 4.1.5 (and doesn't with the modified macro), something has changed in Writer. But FWIW that may also be an improvement, i.e. detecting a mismatch that was disrearded before. BTW, the macro -which just puts a version number in a table in the page footer- works on all versions.
On pc Debian x86-64 with master sources updated today, I don't reproduce this. I also tried with LO Debian package 4.3.3, I don't reproduce this. Terrence/Winfried: could you give a new try with recent LO versions?
(In reply to Julien Nabet from comment #4) > On pc Debian x86-64 with master sources updated today, I don't reproduce > this. > > I also tried with LO Debian package 4.3.3, I don't reproduce this. > > Terrence/Winfried: could you give a new try with recent LO versions? I can't reproduce this on master (source from yesterday afternoon) on openSUSE 13.2, nor on version 4.3.6.2 (openSUSE distribution).
Let's put WFM then. Terrence: don't hesitate to reopen this tracker if you can still reproduce this.
Note to self: try this again when tdf#91702 "Assertion `mnRefCnt>0' failed" at include/vcl/outdev.hxx:342" is fixed.