I think there are a couple of different things going on here.

[These all hold true when running in a text terminal (putty), inside GNU screen
and in an xterm.]

If you try to start wyrd with fewer than 80 columns or 23 lines, you get
assertions to that effect.

If you start wyrd then resize it to fewer than 80 columns or 23 lines, you get
the same assertions.

The thing I have always considered a bug in wyrd is a lot more interesting and
subtle.  If you start wyrd, and _decrease_ the _height_ of its window by
enough lines _in a single resize_ it crashes with the message:

Caught error at toplevel:               │
Invalid_argument(_)

Note that I am not resizing it to fewer than 23 lines.

This doesn't seem to be an issue when changing the number of columns, nor when
increasing, and doesn't seem to be affected by the original number of lines,
or whether or not the number of lines has been increased since wyrd was
started.

To test, start wyrd in an 80x25 xterm, screen, etc.  Increase size to some
large number, then decrease height in lines by a large percent, say 50% --
staying above the minimum requirement of 23 lines - crash.

Now, repeat, but this time, decrease height a few lines at a time, and you
should be able to get all the way down to 23 lines.

I have strace outputs from both scenarios, targeted right around the
nocrash/crash point, and the only thing that looks different is, in the case
where it doesn't crash (smaller decrease in lines) there is something going on
with ignoring SIGTSTP, calling some terminal ioctls, then re-instating the
SIGTSTP handler, but in the case where it _does_ crash, the ignore/re-instate
is not there.  It is over my head, but I can provide the traces if requested.

Thanks,

-Jonathan Hankins

-- 
--------------------------------------------------------------------------
-- Jonathan Hankins    Homewood City Schools                            --
--                                                                      --
-- jhank...@homewood.k12.al.us                                          --
--------------------------------------------------------------------------



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to