Author Topic: Crash on startup  (Read 4490 times)

Offline bughunter2

  • Multiple posting newcomer
  • *
  • Posts: 24
Crash on startup
« on: August 30, 2007, 11:51:54 pm »
Biplab asked me to post this here as well (I posted a bug about this on BerliOS too).

Crash report (RPT file): http://www.nopaste.com/p/aGlm9NTdP

Ok, it (the crash) happened at startup; just before the startup I removed some of the .save, .depend and .layout files. I think this might be causing the crash but I haven't reproduced it.

I hope the crash report is of some help as well as my explanation.

Offline stahta01

  • Lives here!
  • ****
  • Posts: 7591
    • My Best Post
Re: Crash on startup
« Reply #1 on: August 30, 2007, 11:57:48 pm »
What SVN version?
That is about the same error as was happening last week if I recall right.

If compiling C::B yourself, remember to delete the gch files.

Tim S
C Programmer working to learn more about C++ and Git.
On Windows 7 64 bit and Windows 10 64 bit.
--
When in doubt, read the CB WiKi FAQ. http://wiki.codeblocks.org

Offline thomas

  • Administrator
  • Lives here!
  • *****
  • Posts: 3979
Re: Crash on startup
« Reply #2 on: September 01, 2007, 02:09:17 pm »
If compiling C::B yourself, remember to delete the gch files.
This is the proof that the universe was made by a woman.
It is really cynical... now the original code crashes for the same reason the modified code crashed.
The irony lies in the fact that neither the old nor the new code really crashes...
"We should forget about small efficiencies, say about 97% of the time: Premature quotation is the root of public humiliation."

Offline bughunter2

  • Multiple posting newcomer
  • *
  • Posts: 24
Re: Crash on startup
« Reply #3 on: September 01, 2007, 02:14:13 pm »
And yet it does crash? :shock:

Offline thomas

  • Administrator
  • Lives here!
  • *****
  • Posts: 3979
Re: Crash on startup
« Reply #4 on: September 01, 2007, 03:23:23 pm »
The old code has been working without any problems for over 10 months, and the new code (which is now reverted again) effectively differed in one thing:
Code
if(0)
    delete some_pointer;

Technically, it's not if(0) but if(delted), but that variable is initialised to 0 in the class constructor and never changed. Only if you mix revisions, then obviously the old code would not know about initialising that member, which would obviously not work.
This can be by compiling with "old" precompiled headers or by using an "old" plugin with a "new" application, or whatever. Any kind of "old+new" mix is bad.

I've been saying this for almost a week, but people would insist... and they got their will, the old version is back.

Now we have the same identical code that has been working fine for 10 months again, and oh surprise... it crashes in the same location. This proves that the modifications were really harmless.
"We should forget about small efficiencies, say about 97% of the time: Premature quotation is the root of public humiliation."

Offline bughunter2

  • Multiple posting newcomer
  • *
  • Posts: 24
Re: Crash on startup
« Reply #5 on: September 01, 2007, 06:09:17 pm »
Cool, I've deleted the pre-compiled headers, rebuilt C::B, and yet no crashes.