home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.programming:2285 comp.unix.programmer:4257 comp.lang.c:12230 comp.arch:8861
- Path: sparky!uunet!dtix!darwin.sura.net!zaphod.mps.ohio-state.edu!usc!snorkelwacker.mit.edu!ai-lab!fiber-one!sundar
- From: sundar@fiber-one.ai.mit.edu (Sundar Narasimhan)
- Newsgroups: comp.programming,comp.unix.programmer,comp.lang.c,comp.arch
- Subject: Re: What would you like in a debugger?
- Message-ID: <26481@life.ai.mit.edu>
- Date: 12 Aug 92 17:47:52 GMT
- References: <bosullvn.713038630@unix1.tcd.ie> <1992Aug5.193946.16004@linus.mitre.org> <matt.713047463@centerline.com>
- Sender: news@ai.mit.edu
- Reply-To: sundar@ai.mit.edu
- Followup-To: comp.programming
- Organization: MIT Artificial Intelligence Laboratory
- Lines: 27
-
- In article <matt.713047463@centerline.com>, matt@centerline.com (Matt Landau) writes:
- |> In <1992Aug5.193946.16004@linus.mitre.org> hargrove@ramirez.mitre.org (Paul H. Hargrove) writes:
- |> >I want a debugger that works well with X11. This means that if I'm
- |> >tracing an X11 program and it has stopped at a breakpoint, and I
- |> >happen to uncover the window, my X server doesn't just sit there
- |> >trying forever to tell the application to redraw it's window.
- |>
- |> How, pray tell, do you expect your application to redraw its window if
- |> it's stopped at a breakpoint?
- |>
- |> The X server isn't "sit[ting] there trying forever to tell the application"
- |> to do anything at all. The X server generates Expose events for your X11
- |> client application, then forgets about you until something else happens
- |> that you need to know about.
- |>
- |> The fact that your application doesn't read and process these Expose events
- |> when it's stopped at a breakpiont is really not something a debugger can
- |> do anything about. This is kind of the definition of being "stopped at a
- |> breakpoint", is it not?
- I agree with you on this, but the most painful thing about X/Motif
- are programs that involve widgets that do pointer and keyboard grabs.
- I've had breakpoints in callback functions simply hang (freeze up the
- server) because of the grab -- which I can't get out of because to do this
- I would need to TYPE stuff at the debugger telling it to continue.
-
- What is the right thing to do when this happens? (I have a solution
- but it is hardly pretty).
-