<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - cairo-perf-trace failure" href="https://bugs.freedesktop.org/show_bug.cgi?id=73580#c4">Comment # 4</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - cairo-perf-trace failure" href="https://bugs.freedesktop.org/show_bug.cgi?id=73580">bug 73580</a> from <a class="email" href="mailto:emanuele.aina@collabora.com" title="Emanuele Aina <emanuele.aina@collabora.com>"> Emanuele Aina</a> <pre>Testing latest master (4a225fca5f12) on my laptop (Debian sid amd64) I'm still able to reproduce the issue (but debug symbols work everywhere without hacks, yay!): $ CAIRO_TRACE_SO=util/cairo-trace/.libs/libcairo-trace.so util/cairo-trace/cairo-trace zenity --info --text foo cairo-trace: Recording cairo trace data to ./zenity.24699.trace $ perf/cairo-perf-trace ./zenity.24699.trace [ # ] backend test min(s) median(s) stddev. count [ 0] xcb zenity.24699 Error during replay, line 86: invalid value (typically too big) for the size of the input (surface, pattern, etc.) [ 0] xlib zenity.24699 Error during replay, line 86: invalid value (typically too big) for the size of the input (surface, pattern, etc.) [ # ] image: pixman 0.32.6 [ 0] image zenity.24699 Error during replay, line 86: invalid value (typically too big) for the size of the input (surface, pattern, etc.) [ # ] image16: pixman 0.32.6 [ 0] image16 zenity.24699 Error during replay, line 86: invalid value (typically too big) for the size of the input (surface, pattern, etc.) Breaking with gdb on _csi_error() shows that _set_source() is peeking at the stack, expecting a context object and getting a surface object instead. Indeed, the current top of the stack contains: 0 CSI_OBJECT_TYPE_PATTERN 1 CSI_OBJECT_TYPE_SURFACE 2 CSI_OBJECT_TYPE_CONTEXT This is in response to a cairo_set_source_surface() call, so I would suspect that the pattern object is out-of-place but this would mean that there are two bugs: the pattern object in the stack and _set_source() assuming a pattern even when replaying a cairo_set_source_surface() call. That said my understanding of how the cairo script stack management works is too limited to figure out if it's a problem with the trace or with the interpreter, so some help would be greatly appreciated.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the QA Contact for the bug.</li> </ul> </body> </html>