Your suspicion is right. From the same session I've reported the last
crash:
(gdb) break gui_define_fringe_bitmap
Function "gui_define_fringe_bitmap" not defined.
Breakpoint 1 (gui_define_fringe_bitmap) pending.
However, I did follow carefully the instructions. The patch reported
to have
been applied to each of the affected files. And, since I was
unfamiliar with
the procedure, before running `make', I rechecked each of the patched
files to
see if they contained the changed, and they did.
As far as I understand, this means that when I call
`lib-src/emacsclient' it
was actually bringing the installed `emacs' all along. Luckily, for
the
previous tests, it just so happened that it was identical to the
version we
wanted to test. But this is no longer the case now, given the
patch. Is this
correct?
And that's the procedure I had been following given the problem does
not occur
on a standalone Emacs instance.
That given, does starting it with `src/emacs --daemon', then
connecting with
`lib-src/emacsclient' seem like the correct procedure to you?