Re: mvwprintw

[Bob Rossi]

On Fri, Mar 31, 2006 at 01:11:49PM -0500, Thomas Dickey wrote:
> On Fri, 31 Mar 2006, Bob Rossi wrote:
> 
> >>Also, the trace log kind of makes it look like ncurses is doing the
> >>correct thing. What could be the problem?
> >
> >At a minimum, what's the next step I can take in order to find the bug?
> 
> If it's an ncurses bug, then that should be either repeatable by setting
> the same environment, e.g., TERM and the locale stuff, same screen size
> and capturing the output using 'script'.  It simplifies things to 
> eliminate as many variables (and reduce the logsize, of course).

OK, I have a redhat enterprise 3 system where it is repeatable. I'm
sitting on a Windows XP machine. I've logged in via putty, xterm with
cygwin then ssh'd in, and I've even ssh'd in from a ubuntu machine. All
of these terminals produce the error. This makes me think it's not the
shell's fault.

Now, the trace file I sent to you looked good to me. Did it look good to
you? Wouldn't the problem be detectable there? If you didn't see the
file, I'll send it in again in this Email. The line:
(tgdb) b main.c:7
gets corrupted during the last display. Looks like this:
(tgdb) b main1Cc:

Since it looks like ncurses is good, and the terminal is good, could it
be my PTY?

address@hidden ~/cvs/cgdb/cgdb-cvs/cgdb/builddir
$ stty -a
speed 38400 baud; rows 60; columns 80; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>;
eol2 = <undef>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W;
lnext = ^V; flush = ^O; min = 1; time = 0;
-parenb -parodd cs8 -hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl ixon -ixoff
-iuclc -ixany -imaxbel
opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
isig icanon iexten echo echoe echok -echonl -noflsh -xcase -tostop -echoprt
echoctl echoke

> Occasionally I do find some problem using valgrind, e.g., an uninitialized 
> variable.

I tried this again, no bug's found.

> If you make the good/bad configurations "identical", and the trace is the
> the same on both, but the typescript (from script) is different, that 
> still gives some clues, since the trace doesn't cover everything.

Sorry, just want to make sure first. Do you want the "good"
configuration from one computer (ubuntu) running locally, and the "bad"
configuration from the other computer (redhat e3) running via putty?
Would that be acceptable?

Another way it I get a "good" vs "bad" configuration is when I use
libncurses.a (bad) vs libncurses_g.a (good).

> Ultimately all that matters is what goes to the screen.  The trace (for 
> the TRACE_UPDATE flag) shows a picture of the before/after each time a
> refresh is done on a window.  While it's useful, that part can make the 
> trace files very large.
> 
> sorry (was busy).

That's OK! Any help is better than none. As I've said, I'm lost.

Bob Rossi

trace
Description: Text document