[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Status of reported bugs?

From: Christoph Koegl
Subject: Status of reported bugs?
Date: Fri, 18 May 2001 16:07:33 +0200


I reported several bugs in the last few months, and I would like to know if 
there is
any progress/opinion concerning them. Perhaps some of them are fixed/addressed 

On March 1 2001:
In DDD 3.2 I encounter the following behavior very often: I create some displays
while debugging some function; I leave the function or restart the program and
hit some other breakpoint; the displays vanish; I reenter the above function.
Every now and then (I have not yet been able to see the pattern) the displays
do not reappear. In the Displays list these displays are labelled "deferred",
their number has a negative sign. How do I make these displays reappear? Why
does this happen and when? The shown scope is right, so according to the manual
the displays should be (re-)created.

On February 22 2001:
I have a data display of a struct that contains a char-array of fixed length
followed by a pointer. Attempting to double-click on the pointer in order to
pop up a further display results in an error, since the data member labels in
the box look like this:

letter                           = 2 '\002'
inverse                          = "*"
`\000` <repeats 510 times>, next = 0x2d8878

but rather should look like this:

letter  = 2 '\002'
inverse = "*" '\000' <repeats 510 times>
next    = 0x2d8878

Double-clicking on the second line in the original display results in the
following error message (you know what is coming ;-) ):

A parse error in expression, near ` <repeats 510 times>, next)'.

This error (I am using DDD 3.2, didn't try 3.3) should be easy to fix, just
change the line breaking in data displays accordingly.

On January 23 2001:
I propose the following functionality for DDD:

Truncation (optional, perhaps with choice of length) of long strings. This
would help eliminate overly wide (and hence unusable) backtrace etc. windows
in presence of long template instances when debugging C++ programs. My current
backtrace windows is some 4000 pixels wide, the Close button is some estimated
70 cm off the right screen edge. Argh.
I also noticed that gdb outputs (sometimes) very long frame annotations in
response to up/down commands. These seem to slow down DDD's execution speed
very much. I don't know if gdb does this by default (since I do not use vanilla
gdb very often these days) or if DDD sets up custom frame annotations.
Very long strings also greatly diminish the utility of the data window.
Especially when using the STL in C++ programs one always has nested template
instances that result in veeeery long instance names. The boxes in the data
window are sometimes thousands of pixels wide, navigation is, well, suboptimal.
I think a truncation style such as StartingChars...EndingChars with user
defined values for the number of starting and ending characters would be nice.
In the data window I would like to see a popup menu entry "Truncate..." that
enables the user to truncate the string to the left of some "label = thing"
display entry.

What do you think?

Cheers, Christoph

PS. Another question: One of my applications is 60 MB with debug information.
    Gdb needs some time to read in the symbols table etc. But with DDD the
    loading takes three times as long (at least). Why? What kind of initiali-
    zation takes place within DDD? I just hate waiting 3 Minutes (on a Sparc
    Ultra 2 with 1G of RAM) for DDD to fire up ... ;-)

Bye, Christoph

reply via email to

[Prev in Thread] Current Thread [Next in Thread]