[comp.windows.x] Sun OW2.0 mailtool crashes MIT server

randy@erik.UUCP (Randy Brown) (05/17/91)

This isn't a formal bug report, so don't complain that I didn't follow
the standard form; it's a request to see if anyone else has seen this or
a similar problem to let me know if it's worth chasing. 

Environment: Sun 4/40 (IPC), SunOS 4.1
	     MIT X11R4 (patch level 18) server
	     window manager: OpenWindows 2.0 olwm << OR >> X11R4(18) twm
	     compilation of X11R4 with Sun's bundled cc

The client /usr/openwin/bin/xview/mailtool, in the environments
producing either ldd output

  -lxview.3 => /usr/openwin/lib/libxview.so.3.0
   -lolgx.3 => /usr/openwin/lib/libolgx.so.3.0
-ldeskset.0 => /usr/openwin/lib/libdeskset.so.0.0
    -lX11.4 => /usr/openwin/lib/libX11.so.4.3 <<OR>> /usr/lib/libX11.so.4.2 
      -lc.1 => /usr/lib/libc.so.1.5

sometimes crashes the server.  The problem occurs when deleting the
message at the end of the list of messages.  It most often occurs when
there are only one or two other message headings displayed.  It may be
occurring during an attempt by mailtool to scroll new message headings
into view.  My first wild guess was that there may be some screwup with
a shared pixmap, but ipcs reports no shared memory segments in use.  (Of
course, ipcs may be broken.)

It looks as if the window containing the message headings is extended to
the bottom of the screen (overwriting all other windows, including the
window manager's decorations).   From dbx:

    program terminated by signal SEGV (no mapping at the fault address)
    (dbx) where
    cfbFillBoxSolid() at 0x10e44
    cfbPaintWindow() at 0x7e34
    miSpritePaintWindowBackground() at 0x925c0
    miHandleExposures() at 0x88d08
    cfbCopyArea() at 0xc1d8
    miSpriteCopyArea() at 0x933ac
    ProcCopyArea() at 0x2bb98
    Dispatch() at 0x29300
    main() at 0x3ae04

Does anyone else see this?  (I hope it's not a FAQ... I checked the file
and the answer didn't jump out at me.)  Any clues?  Unfortunately, the
suggestion "don't use Sun deskset apps" isn't helpful; we can't advise
our customers to use the MIT server if it is likely to crash on a
commonplace client, even if it isn't our client.  (We ship the computer
and software as a small part of a large system, and they tend to hold us
responsible for the darndest things :-)

Randy Brown (uunet!erik!rbrown)

bill@polygen.uucp (Bill Poitras) (05/20/91)

In article <9105171539.AA03425@erik.uucp> randy@erik.UUCP (Randy Brown) writes:
>Does anyone else see this?  (I hope it's not a FAQ... I checked the file
>and the answer didn't jump out at me.)  
Yes I have seen this.  My work around, only delete messages from the front
of the list.

+-----------------+---------------------------+-----------------------------+
| Bill Poitras    | Polygen Corporation       | {princeton mit-eddie        |
|     (bill)      | Waltham, MA USA           |  bu sunne}!polygen!bill     |
|                 | FAX (617)890-8694         | bill@polygen.com            |
+-----------------+---------------------------+-----------------------------+