Hi there, I just commited the code that cleanups the bookmarks server. It works like this: <q> /* * If this is our first run, kill the (old) bookmarks server, * and create the '~/.dillo/.072' file to flag what we did. */ static void Dillo_check_bm_srv12(char *dir) [...] </q> Yes, it uses the "[b]m_srv12" trick and some other precautions. Please test it: the whole point is that when installing the new version of dillo (072), the old bookmarks server may keep running and survive the updated. The new code detects the first dillo-072 run and issues a kill command to "bm_srv12". Onto another topic: A last minute leak fix (iterators) left dillo a bit unstable. We knew the iterators had a leak and it's great to have found it. Now it will take some extra time to try to fix the problem and, after that, stress-test the Copy&Paste code again. We haven't got an accurate procedure to reproduce the crash (yet). This explanation may help: <q> I think I have found a relatively new segfault bug : You need some sort of thumbtags like the screenshots on http://www.incagold.com/products/Speed/speedthief.php Now you have to move down the mouse and at the same time click on one image (i.e. click while still aiming for the image). If the button-push occures outside the image and the button-release inside the image, chances are high to get a segfault. This is a bit hard to reproduce, but I couldn't get an older version with selection to behave like that. It also happened a few times to me last night (after the last 'cvs up') but never before. </q> We'll be working on it. Cheers Jorge.-
Onto another topic:
A last minute leak fix (iterators) left dillo a bit unstable. We knew the iterators had a leak and it's great to have found it. Now it will take some extra time to try to fix the problem and, after that, stress-test the Copy&Paste code again.
We haven't got an accurate procedure to reproduce the crash (yet). This explanation may help:
Here is an alternative method : simply RIGHT-click on the thumbnails in said page : http://www.incagold.com/products/Speed/speedthief.php Curiously, on FreeBSD I "only" get a Gtk-WARNING **: gtk_signal_emit_by_name(): could not find signal "link_pressed" in the `DwTable' class ancestry a second attempt to right click the image will pop up the menu. Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198
Andreas Schweitzer writes:
We haven't got an accurate procedure to reproduce the crash (yet). This explanation may help:
Here is an alternative method : simply RIGHT-click on the thumbnails in said page : http://www.incagold.com/products/Speed/speedthief.php
Curiously, on FreeBSD I "only" get a Gtk-WARNING **: gtk_signal_emit_by_name(): could not find signal "link_pressed" in the `DwTable' class ancestry a second attempt to right click the image will pop up the menu.
For what it's worth, on OpenBSD with Dillo from CVS about a week old, I get no warnings and no crashes when I try this test. -- Chris music is what numbers feel like San Francisco, CA http://www.nodewarrior.org/chris
On 2003-04-24 at 20:31 +0200, Andreas Schweitzer wrote:
Here is an alternative method : simply RIGHT-click on the thumbnails in said page : http://www.incagold.com/products/Speed/speedthief.php
Ah, that did it, I was finally able to reproduce.
Curiously, on FreeBSD I "only" get a Gtk-WARNING **: gtk_signal_emit_by_name(): could not find signal "link_pressed" in the `DwTable' class ancestry a second attempt to right click the image will pop up the menu.
FreeBSD 4.7, x86, dumped core on sigsegv. Stack backtrace: -----------------------------< cut here >------------------------------- #0 0x283df1d6 in strncpy () from /usr/lib/libc_r.so.4 #1 0x28273f0e in g_strndup () from /usr/local/lib/libglib12.so.3 #2 0x80671e9 in Selection_copy (selection=0x80eb560) at selection.c:379 #3 0x8066dfc in a_Selection_button_release (selection=0x80eb560, it=0x8281500, char_pos=0, link=39, event=0x80f6018, within_content=1) at selection.c:202 #4 0x805f916 in Dw_page_send_selection_event (page=0x8123e00, fn=0x8066cb8 <a_Selection_button_release>, x=178, y=320, event=0x80f6018) at dw_page.c:1486 #5 0x805f98b in Dw_page_button_release (widget=0x8123e00, x=178, y=320, event=0x80f6018) at dw_page.c:1513 #6 0x80651c3 in Dw_widget_mouse_event (widget=0x811b700, viewwidget=0x80c0b00, x=283, y=1175, event=0x80f6018) at dw_widget.c:740 #7 0x805b86c in Dw_gtk_viewport_mouse_event (widget=0x80c0b00, x=283, y=330, event=0x80f6018) at dw_gtk_viewport.c:358 #8 0x805b940 in Dw_gtk_viewport_button_release (widget=0x80c0b00, event=0x80f6018) at dw_gtk_viewport.c:389 #9 0x2818ce93 in gtk_marshal_BOOL__POINTER () from /usr/X11R6/lib/libgtk12.so.2 #10 0x281bced6 in gtk_signal_real_emit () from /usr/X11R6/lib/libgtk12.so.2 #11 0x281baef3 in gtk_signal_emit () from /usr/X11R6/lib/libgtk12.so.2 #12 0x281f2f6c in gtk_widget_event () from /usr/X11R6/lib/libgtk12.so.2 #13 0x2818cde5 in gtk_propagate_event () from /usr/X11R6/lib/libgtk12.so.2 #14 0x2818bf56 in gtk_main_do_event () from /usr/X11R6/lib/libgtk12.so.2 #15 0x2823c87c in gdk_event_dispatch () from /usr/X11R6/lib/libgdk12.so.2 #16 0x2826acf3 in g_main_dispatch () from /usr/local/lib/libglib12.so.3 #17 0x2826b31c in g_main_iterate () from /usr/local/lib/libglib12.so.3 #18 0x2826b4b4 in g_main_run () from /usr/local/lib/libglib12.so.3 #19 0x2818b813 in gtk_main () from /usr/X11R6/lib/libgtk12.so.2 #20 0x8067e25 in main () #21 0x8053269 in _start () -----------------------------< cut here >------------------------------- At level #2, g_strndup() is called: 379 else if (cmp == 0) 380 /* the end */ 381 txt = g_strndup(si->content.data.text, 382 b_char); (gdb) print si->content.data.text $4 = 0x81265b0 "3D" (gdb) print b_char $5 = 0 Are you supposed to be able to dup zero-length of a string?
participants (4)
-
Andreas Schweitzer -
Chris Palmer -
Jorge Arellano Cid -
Phil Pennock