On Fri, 28 May 2004 22:46, Rene Engelhard wrote: > > What do you suggest? Disabling this feature for Linux/sparc only and > enabling it for the rest?
i think you should not --enable-crashdump for any platforms, as you would need a server to receive the messages such as shown here: http://qa.openoffice.org/issues/show_bug.cgi?id=17635 #ifdef OSL_PRODUCT - #if defined(BUILD_SPECIAL) || defined(OOO_RELEASE_BUILD) + #if defined(ENABLE_CRASHDUMP) Value = "report.services.openoffice.org"; #else point your browser at that link to see there is some infrastructure: report.services.openoffice.org it is no good sending to that one, it will go into the bit bucket, so you will need add to the codes in sal/osl/unx the backtrace looks like this, so the user would need copy and paste into reportbug. you can put a text message where it says "Stack:" in signal.c line 617 : line 78: /home/jim/OpenOffice.org1.1.2/program/crash_report_with_gtk.bin: No such file or directory Fatal exception: Signal 10 Stack: /home/jim/OpenOffice.org1.1.2/program/libsal.so.3+0x25eb4[0x70ce9eb4] /home/jim/OpenOffice.org1.1.2/program/libsal.so.3+0x25fc4[0x70ce9fc4] /lib/libpthread.so.0+0xa89c[0x7130e89c] /lib/libc.so.6+0x32f98[0x715baf98] /home/jim/OpenOffice.org1.1.2/program/libtl645ls.so+0x89130 (_ZN12SvFileStream7GetDataEPvm+0x24)[0x70a1d130] /home/jim/OpenOffice.org1.1.2/program/libtl645ls.so+0x8632c(_ZN8SvStream4ReadEPvm+0x1a4) [0x70a1a32c] <snip> the behaviour is controlled from signal.c where it has #ifdef ENABLE_CRASHDUMP, also #ifdef LINUX,
