bug#5851: 24.0.50; first character after opening quote often gets eaten in CC modes

[Alan Mackenzie]

Hello, Stefan.

On Sun, Nov 24, 2019 at 00:42:12 +0100, Stefan Kangas wrote:
> Stefan Kangas <stefan@marxist.se> writes:

> > npostavs@gmail.com writes:

> >> Paul Pogonyshev <pogonyshev@gmx.net> writes:

> >>> FWIW the bug is still present.  I get this backtrace (2011-02-19
> >>> build):

> >>>     (error "Invalid search bound (wrong side of point)")
> >>>       signal(error ("Invalid search bound (wrong side of point)"))
> >>>       byte-code(" b\210\302    @    A\"\207" [start err signal] 3)
> >>>       c-syntactic-re-search-forward("[;{}]" 8510 end)
> >>>       c-before-change-check-<>-operators(6462 6462)
> >>>       #[(fn) "     \n\"\207" [fn beg end] 
> >>> 3](c-before-change-check-<>-operators)
> >>>       mapc(#[(fn) "     \n\"\207" [fn beg end] 3] 
> >>> (c-extend-region-for-CPP c-before-change-check-<>-operators))
> >>>       c-before-change(6462 6462)
> >>>       self-insert-command(1)
> >>>       call-interactively(self-insert-command nil nil)

> >>> I cannot find a reliable way to reproduce, but it happens often
> >>> enough.  Apparently, buffers not touched in quite a while are more
> >>> prone.  I.e. open a C++ file and leave it alone for 5 minutes.  Then
> >>> switch to the buffer and type outside any string a double quote
> >>> following by any other character --- the second character will likely
> >>> get eaten with an error as above.  M-x toggle-debug-on-error helps
> >>> spotting it.

> >> Maybe this is fixed in master now, along with Bug#28850?  (the error
> >> there is the same, though the backtrace is a bit different)

> > Indeed, it looks like this has been fixed on master -- or at least I
> > can't reproduce it.  Alan, what do you think?

> No further comments within 6 weeks.  I'll give it a couple of more
> weeks, and if no one objects within that time, I'll just assume that
> this has indeed been fixed and close this bug report.

I think the time has come to close the bug.  I think it highly likely
it's been fixed, probably quite some while ago.  There have been quite a
few bugs with similar symptoms over the years.

> Meanwhile, if anyone has been seeing this issue, please try it on
> current master to see if the issue persists, then report back your
> findings.

> Best regards,
> Stefan Kangas

-- 
Alan Mackenzie (Nuremberg, Germany).