===================0.5.3 release=========================== 2008-05-25 Dodji Seketeli * src/dbgengine/nmv-gdb-engine.cc, nmv-i-debugger: Added the IDebugger::step_instruction() call to step into the next machine instruction, instead of stepping into the next C instruction. * src/persp/dbgperspective/nmv-dbg-perspective.cc: (DBGPerspective::on_debugger_command_done_signal): step into the next machine instruction instead of over the next instruction. This might help in fixing bug #534695 2008-05-25 Dodji Seketeli * src/dbgengine/nmv-gdb-engine.cc: (in the OnCommandDoneHandler): don't set the engine state to IDebugger::READY when the command that has been DONE is "detach-from-target". This should fix #534619 2008-05-18 Dodji Seketeli * src/dbgengine/nmv-gdb-engine.cc: emit the IDebugger::detached_from_target_signal() when the inferior program exits. * src/persp/dbgperspective/nmv-dbg-perspective.cc: (DBGPerspective::on_attached_to_target_signal (bool a_is_ready)): when the debugger detaches from the inferior, update the different sentivity related menu states. Also stop the throbber and the busy cursor. (DBGPerspective::on_program_finished_signal()): make sure to notify the world that we detached from target. Also make sure in this case to "de-sensitivate" the menu items accessible when the debugger is connected to the target. (DBGPerspective::on_debugger_state_changed_signal()): don't act upon states that are different from IDebugger::READY because those are notified to client code via proper signals and are thus taken in account elsewhere. Don't do more work than needed. * src/persp/dbgperspective/nmv-memory-view.cc: (on_debugger_stopped): don't try to query the debugger when we it the inferior is exiting. This leads to further errors and and will later change the state of the debugger, likely bringing it to a IDebugger::READY state, for instance. At some point the current implementation of IDebugger must do what is necessary to to not switch to the IDebugger::READY state in those erratic cases. For now, let's just behave gently. * src/persp/dbgperspective/nmv-registers-view.cc: same thing as above. This entry should fix bug #482806. 2008-05-18 Dodji Seketeli * src/dbgengine/nmv-dbg-common.h, src/dbgengine/nmv-gdb-engine.[cc,h], src/dbgengine/nmv-gdbmi-parser.cc, src/dbgengine/nmv-i-debugger.h, src/persp/dbgperspective/nmv-call-stack.cc, src/persp/dbgperspective/nmv-dbg-perspective.cc, src/persp/dbgperspective/nmv-local-vars-inspector.cc,: Replace stop reasons string use by the enum StopReason use. Move StopReason from the OutOfBandRecord type to the IDebugger interface type. That way, it can be used from within clients of IDebugger. Change the code of the client code accordingly. 2008-05-12 Dodji Seketeli * src/common/Makefile.am, src/confmgr/Makefile.am, src/dbdimpl/sqlite/Makefile.am, src/dbgengine/Makefile.am, src/persp/dbgperspective/Makefile.am, src/uicommon/Makefile.am, src/workbench/Makefile.am: added the -Wl,--as-needed link flags to remove useless libraries from the .dynamic section of the DSOs. 2008-05-11 Dodji Seketeli * src/common/nmv-env.cc,h: added build_path_to_help_file() entry point to compute the path the a nemiver help file. That is needed to avoid using gnome_help_display() and thus libgnome. * tests/test-env.cc: added this test file for the new entry point. * src/workbench/nmv-workbench.cc: (Workbench::on_contents_menu_item_action): don't use gnome_help_display(). Instead, spawn yelp with the path to to the nemiver help file. Okay this is a bit rude but there is no better way right now. * configure.ac: don't detect libgnome anymore. * src/main.cc: don't initialize libgnome anymore. 2008-05-06 Dodji Seketeli * src/uicommon/nmv-terminal.cc: _GNU_SOURCE is defined on freebsd as well, so use don't use it to discriminate that platform. This should hopefully close #529718 2008-05-05 Dodji Seketeli * src/uicommon/nmv-terminal.cc: include the right headers in lieu of pty.h, on FreeBSD. Patch from Romain Tartière 2008-05-05 Dodji Seketeli * src/dbgengine/nmv-gdb-engine.cc: don't include pty.h It is useless here and breaks FreeBSD builds. Patch from Romain Tartière 2008-05-05 Dodji Seketeli * src/common/nmv-ustring.cc: on FreeBSD, define strnlen(). Patch from Romain Tartière . 2008-05-05 Dodji Seketeli * src/common/nmv-proc-utils.cc: Define __MAX_BAUD for FreeBSD. Include specific FreeBSD headers for process manipulation. Patch from Romain Tartière 2008-05-05 Dodji Seketeli * src/common/nmv-log-stream-utils.h: For freebsd, make sure __ASSERT_FUNCTION is defined. Patch from Romain Tartière . 2008-05-04 Dodji Seketeli * data/nemiver.desktop.in: remove obsolete fields. A courtesy from Luca Bruno 2008-05-03 Dodji Seketeli * src/dbgengine/nmv-gdbmi-parser.h: don't throw exception when trying to read empty lists of result or values. This should hopefully fix bug #526863. 2008-04-12 Dodji Seketeli * src/main.cc: reorganised the command line options parsing. Added init_option_context() to - guess what - initialize the option context and add the nemiver and gtk+ groups to it, so that both gtk+ and nemiver options can be available. That way, both nemiver and gtk+ help messages can be displayed. Make sure parse_command_line() does initialize the option context by calling the brand new init_option_context(), before actually trying to parse the command line. This should fix #527696 – nemiver --help only shows libgnome related help messages 2008-04-11 Dodji Seketeli * configure.ac: make sure to AC_SUBST the deps versions number variables so that we can use them in foo.in files like *.pc.in files. * src/common/libnemivercommon.pc.in: make sure to add ligtop to the list of deps in the generated libnemivercommon.pc file. This should fix bug #526826 – missing dependency in .pc file 2008-04-11 Dodji Seketeli * src/dbgengine/nmv-gdbmi-parser.cc: fix a parsing bug that happens when setting a pending breakpoint on a fully qualified function. * tests/test-gdbmi.cc: add a regression test for this bugfix. This entry should fix bug #526866 2008-04-08 Jonathon Jongsma * src/persp/dbgperspective/nmv-memory-view.cc: properly set memory editor widget to insensitive when the debugger is running. Otherwise the user could change the memory values and it would attempt to send those values to gdb when it wasn't ready for them. 2008-04-08 Dodji Seketeli * src/dbgengine/nmv-i-debugger.h: add IDebugger::is_attached_to_target() to let client know if IDebugger is attached to a target or not. * src/dbgengine/nmv-gdb-engine.cc: basically track when we are attached to a target or not. * src/persp/dbgperspective/nmv-dbg-perspective.cc: When we detach from a target, the "detach from target" menu should be set to insensitive. This should fix #526696. 2008-04-07 Dodji Seketeli * README: updated.