|
From: Dan K. <da...@ke...> - 2003-07-24 05:14:46
|
What does ==21616== Invalid memory access of size 2 ==21616== at 0x4794C81B: XMLShapeStyleContext::FillPropertySet(com::sun::star::uno::Reference<com::sun::star::beans::XPropertySet> const&) (in /opt/OpenOffice.org1.1rc1/program/libxo645li.so) mean? Was it a read access or a write access? - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045 |
|
From: Julian S. <js...@ac...> - 2003-07-24 07:57:48
|
On Thursday 24 July 2003 06:18, Dan Kegel wrote: > What does > ==21616== Invalid memory access of size 2 > ==21616== at 0x4794C81B: > XMLShapeStyleContext::FillPropertySet(com::sun::star::uno::Reference<com::s >un::star::beans::XPropertySet> const&) (in > /opt/OpenOffice.org1.1rc1/program/libxo645li.so) mean? Was it a read > access or a write access? > - Dan Addrcheck won't tell you, because I was too lazy to write different but basically identical code to deal with read vs write accesses. Use --skin=memcheck to find out, since that does distinguish reads from writes. J |
|
From: Dan K. <da...@ke...> - 2003-07-24 13:39:19
|
Julian Seward wrote: > On Thursday 24 July 2003 06:18, Dan Kegel wrote: > >>What does >>==21616== Invalid memory access of size 2 >>==21616== at 0x4794C81B: >>XMLShapeStyleContext::FillPropertySet(com::sun::star::uno::Reference<com::s >>un::star::beans::XPropertySet> const&) (in >>/opt/OpenOffice.org1.1rc1/program/libxo645li.so) mean? Was it a read >>access or a write access? > > Addrcheck won't tell you, because I was too lazy to write different > but basically identical code to deal with read vs write accesses. > Use --skin=memcheck to find out, since that does distinguish reads > from writes. I can't use --skin=memcheck, as OpenOffice has so many uninitialized memory references Valgrind stops reporting errors. IMHO it would be worth it to make addrcheck distinguish between reads and writes. The problem I'm facing is convincing the chowderheads that a Valgrind warning is a serious bug. It'd be easier if the warning said the access was a write. - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045 |
|
From: Steve G <lin...@ya...> - 2003-07-24 13:48:03
|
>The problem I'm facing is convincing the chowderheads >that a Valgrind warning is a serious bug. That's a shame. Valgrind testing and subsequent bug fixing made the freeradius 0.9.0 release this week stable. Without all the bug fixing pointed out by valgrind, I'm sure their mail list would have been eaten up by user bug reports. I've also wondered the same thing about whether the access was a read or write. It would make bug fixing a little faster. Sometimes I notice that valgrind is 1 or 2 lines off when it mentions the line number. Having a read or write message would help make a positive identification of the guilty code when line numbers are out of sync. -Steve Grubb __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com |
|
From: Dan K. <da...@ke...> - 2003-09-05 23:35:13
|
Nicholas Nethercote wrote: > On Thu, 24 Jul 2003, Dan Kegel wrote: > >>I can't use --skin=memcheck, as OpenOffice has so many uninitialized >>memory references Valgrind stops reporting errors. >> >>IMHO it would be worth it to make addrcheck distinguish >>between reads and writes. The problem I'm facing is >>convincing the chowderheads that a Valgrind warning is >>a serious bug. It'd be easier if the warning said >>the access was a write. > > I've just committed to the HEAD a change that does this. Bless you. - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045 |