I wrote:
> As expected, booting into kernel 5.7.17-200.fc32.x86_64 allows me to run
> poplog as usual, without any problems.
I was wrong -- the previously reported 'access violation' message + Setpop
some time after starting ved is still there.
It seems to be triggered when 'vedautosave' first runs. It's a minor
nuisance, because I can then run
: ved
and get back to where I was in the editor.
After that, ved works for long periods without any problem (e.g. when
working on a paper I can sometimes keep the same ved process going for
several days, persisting through hibernate + restore).
I don't normally use xved (I find multiple windows unnecessary for most of
my use of ved -- e.g. reading and writing email messages and writing
papers). But I have checked that the 'access violation' message also
occurs (once) when running Xved but the Xved window is left untouched: I
then don't have to redo the 'ved' command to continue editing, as I have to
do if not running xved when the error is reported.
If I move the XVed window out of the way I see this:
<<<<<<< Access Violation: PC = 00007FFFF7826890, Addr = 00000007FF618C68,
Code
= 1 >>>>>>>
;;; MISHAP - serr: MEMORY ACCESS VIOLATION (see above)
;;; PRINT DOING
;;; DOING : <External Calls>
----
I have just tried triggering that while running gdb (with ved, not xved),
and got the usual error message. but it seems gdb needs help:
Program received signal SIGSEGV, Segmentation fault.
tl( [] ) => musable
(mem=0x608e70 <pop_mem_block+20080>) at malloc.c:4855
4855 else if (inuse (p))
Missing separate debuginfos, use: dnf debuginfo-install
libX11-1.6.12-1.fc32.x86_64
(gdb)
However, if pop11 is trying to apply 'tl' to an empty list when autosave is
first triggered that may be fixable by an extra test in the code. Later
when I have caught up with other things I'll see if I can work out what has
happened.
So far I've found that 'inuse' is not an identifier available at runtime in
poplog, and after searching through a lot of Poplog system files using grep
I can't find an occurrence, so I assume the line 4855 comes from some debugging
code in gdb ??
It looks as if I can't get further without invoking
dnf debuginfo-install libX11-1.6.12-1.fc32.x86_64
But I must first try to catch up with other urgent things first.
Aaron
