From: Craig Chambers (chambers@cs.washington.edu)
Date: Tue Feb 11 2003 - 15:15:38 PST
Andrei Alexandrescu wrote:
>
> It is quite pleasurable to work and debug code in Vortex and Whirlwind. I
> have a couple of suggestions that would improve my experience quite some:
>
> 1. If the command line had history, my productivity would blow the roof
> away. It doesn't, nor it has the most basic editing features on my system
> (backspace doesn't work (I need to press ctrl-bcksp), arrows don't work,
> etc.) Plus, a short mnemonic such as Ctrl-R "Repeat the last command" would
> be great instead of typing next next next... etc. when debugging.
You can run Vortex under emacs or something, and get this kind of thing for
free. There's also a library for stdin reading that supports all that fun
editing you describe; maybe we should link it in.
>
> 2. Next, step, etc. should have shorter mnemonics.
But we don't want to clash with likely variable names. "n" and "s" would be
bad.
What I do, if I'm going to repeat a command, is copy it into my X cut buffer,
and then just paste it over and over.
>
> 3. I'd like a command that just kills the process being debugged
> immediately. Now I have to clear breakpoints and run to completion.
> Awkward.
^C? really_quit?
>
> 4. The "type 'help' for help" added line was very useful exactly for the
> first time I used the debugger. After that it became unnecessary clutter.
But a small amount, and if you miss it the first time....
>
> 5. A 'nextline' command that breaks not to the next command, but to the
> next line in the debugger would be great.
Hmm, interesting. There might in fact be enough debug info available to do
this. It's like next, except that if the next goes to the same source line as
before, then keep going, until the source lines differ.
>
> 6. This would be fantastic: I want to 'next' all message sends that belong
> to other files and 'step' into all message sends that belong to the file
> I'm in. A command like 'filestep' would be wonderful.
Hmm. This might also be feasible.
>
> As Craig said, these are just things that I could do myself, so this is a
> note for my future self (when I'll hopefully have more time... such as
> after I finish writing my thesis many, many years from now) as much as a
> request for opinions on whether you consider these worthwhile additions.
>
> Andrei
>
> _______________________________________________
> Cecil mailing list
> Cecil@cs.washington.edu
> http://mailman.cs.washington.edu/mailman/listinfo/cecil
-- Craig
_______________________________________________
Cecil mailing list
Cecil@cs.washington.edu
http://mailman.cs.washington.edu/mailman/listinfo/cecil
This archive was generated by hypermail 2.1.5 : Tue Feb 11 2003 - 15:15:45 PST