Re: emacs oddity

[markus.hoenicka@mhoenicka.de]

Markus Hoenicka writes:
 > Am 2014-05-21 13:46, schrieb Ken Brown:
 > > On 5/21/2014 3:13 AM, Markus Hoenicka wrote:
 > >> At 2014-05-21 03:42, Ken Brown was heard to say:
 > >>> On 5/20/2014 7:57 PM, Max Polk wrote:
 > >>>> I've been noticing the background timer giving odd messages from
 > >>>> emacs-w32 once in a while such as the following when not touching 
 > >>>> the
 > >>>> keyboard and the editor is idle.:
 > >>>> 
 > >>>> seconds-to-time: Arithmetic range error: "round", -0.0e+NaN
 > >>>> seconds-to-time: Arithmetic range error: "round", 0.0
 > >>>> 
 > >>>> The stack dump of a recent emacs crash report on this list seemed to
 > >>>> be in background timer code.  Maybe we found something.
 > >>> 
 > >>> Have you tried the test release
 > >>> (https://cygwin.com/ml/cygwin/2014-05/msg00047.html)?
 > >>> 
 > >> 
 > >> Hi,
 > >> 
 > >> I haven't followed this thread in all its gory details, but I recall
 > >> that only emacs-w32 was reported to have problems. I have noticed 
 > >> random
 > >> crashes and weird messages without prior user actions in emacs-x11 
 > >> since
 > >> I moved from WinXP + Cygwin x86 to Win7 + Cygwin x86_64 a few weeks 
 > >> ago.
 > >> None of these problems ever occurred on either FreeBSD (amd64, Emacs 
 > >> 24)
 > >> or Debian (amd64, Emacs 23). I just thought I'd mention this because 
 > >> the
 > >> problem does not seem to be limited to the -w32 build of Emacs. I've
 > >> preserved the odd messages on three separate occasions:
 > >> 
 > >> Args out of range: [t 21335 39727 373923 0.5 
 > >> blink-cursor-timer-function
 > >> nil nil 100000], 4
 > >> 
 > >> Invalid function: #[(timer) "^H >^H
 > >>> ^[\211^203^Q^@Å^H        \"^Q^K\203^Z^@Å^H
 > >> \"^R^L\206^_^@^K*\207" [timer timer-list timer-idle-list cell2 cell1
 > >> delq] 4 2245674]
 > >> 
 > >> timer-relative-time: Wrong type argument: vectorp, [t1 time high low
 > >> micro pico nil 3 0 2 ...]
 > >> 
 > >> All of them indicate timer related problems as Max pointed out. I do 
 > >> not
 > >> have meaningful backtraces of the crashes at this time because the
 > >> installed emacs-x11 seems to lack debug symbols.
 > > 
 > > You have to install emacs-debuginfo in order to get the symbols.
 > > 
 > 
 > Thanks, I've installed it now just in case. Until now, no crashes or 
 > weird messages with emacs-X11-24.3.90-1 yet...
 > 
 > regards,
 > Markus
 > 

Hi,

I've been testing emacs-X11-24.3.90-1 for more than two weeks now, and
this is what happened:

- a few hours after I claimed that there were no problems with this
  version yet (see above), Emacs crashed again.

- I used to start Emacs from gdb for about two weeks, hoping to catch it
  red-handed. No problems occurred

- I reverted to starting Emacs as I had done for years (through
  startxwin), but I've set up the CYGWIN environment variable to
  create true core dumps. A couple of days later Emacs crashed again
  and left a core dump. I've sent that along with a problem
  description to bug-gnu-emacs AT gnu.org:
  http://debbugs.gnu.org/cgi/bugreport.cgi?bug=17753

This basically means that the current test release is a major
improvement although it did not eliminate crashes altogether. As I do
not have any core dumps of the previous Emacs version, I can't even
tell whether or not the same bugs are involved.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple