This is the mail archive of the cygwin mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Self-destructing Insight installation!


    Hi everyone,

  So there I was, debugging the gnu cpp with insight, then making a few
changes, rebuilding cpp0.exe and doing another debug run.

  And all of a sudden, I got this error message, and insight refused to run.

  I've only managed to capture it in .PNG format, because it was one of
those windows that you can't copy+paste code from, but basically what it
said was "Can't find a usable main.tcl", then an error report ("extra
characters after close-brace") and stack backtrace, followed by the
threatening message "This probably means that insight wasn't installed
properly".  Which was surprising, because I hadn't been messing around with
my cygwin installation any during this time.

  Anyways, what transpires is that the crash was occurring while reading the
prefs for the application, and during my previous run of insight I had set a
breakpoint.  Somehow, when this got written out into my local prefs file
(~/gdbtk.ini), it was written in a malformed syntax  that caused the error
next time I tried running insight when it went to source the prefs.

  This is absolutely repeatable: I can add or remove the line either with a
text editor or (when insight is working) by using the break command in the
console window, and when the line is present insight won't start, and when
it's gone it will start.

  Here's how the breakpoints line was being written out into my prefs file:

---snip---
	C:/artimi.src/gcc-build/gcc/gcc/cpp0.exe/breakpoints={{break
cpp_post_options} 1 {} {}} {{break main} 1 {} {}}
---snip---

  I haven't pursued this any further.  If anyone wants to see the .PNG file
of the error message, or the rest of my gdbtk.ini, I could mail it to them.
However, I imagine that there's probably a very simple bug in the routine
that writes out the prefs, and I hope this report is enough that someone
who's more familiar with the insight code might be able to just point right
at it and go "There it is!".  If not, just let me know what further
diagnostics you'd like.

  Here's my insight and tcl version info, from "insight --version" and
"runtest --version"

GNU gdb 2003-09-20-cvs (cygwin-special)
Tcl version is          8.4

  Of course, maybe this is already solved in 8.4.1; I haven't tried a
snapshot, but I couldn't find anything that sounds like this problem when I
searched the insight ml archive.

    cheers, 
      DaveK
-- 
Can't think of a witty .sigline today....


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]