Are crash reports useful

Eric Chatonet eric.chatonet at sosmartsoftware.com
Thu Aug 7 12:19:20 EDT 2008


Jacque,

As I said I'm sure that Runrev has nothing to make with our code but  
only with 'exceptions' in the engine.
But it would be clearer to add a text file in the logs folder to  
explain all this :-)

Le 7 août 08 à 18:09, J. Landman Gay a écrit :

> Eric Chatonet wrote:
>
>> I let you report this thread to Mark W. ;-)
>
> I asked Mark to respond, and he asked me to post to this list, as  
> he isn't subscribed here. Basically, it is true that the format of  
> the logs is generated by the OS and there is nothing to worry about  
> regarding confidentiality. Here is what he said:
>
> ***
> The crash logs are those generated by the 'MiniDump' component that  
> is present in Windows. This is a system provided library that  
> allows applications to produce a dump containing the state of the  
> call stack when it crashes. It is the same component that is used  
> by the default Microsoft handler for these which usually prompts  
> you to ask if you 'want to send information to Microsoft to  
> diagnose'. As far as I'm aware this information is entirely  
> unencrypted - its just encoded in a binary form.
>
> The information present in the file is pretty much identical to the  
> information you get when a CrashLog from CrashReporter on OS X.  
> Although different settings (small, medium, large) can result in  
> more information being included - which means they can be closer to  
> UNIX-style 'Core dumps' than a simple log. As far as I'm aware,  
> 'small' contains just a call stack, 'medium' contains the call- 
> stack and contents of variables on the stack and 'large' contains  
> any interesting segments of memory (although I've never been able  
> to use 'large' dumps to any more effect than 'small' or 'medium').
>
> Rather than a text file, the minidump file is binary file which has  
> the advantage of being loadable directly into Visual Studio for  
> post-mortem analysis. Using this (by combining with special debug  
> symbol files we keep from compiling) we can typically jump directly  
> to the line that caused Revolution to crash - as you can imagine  
> this can make determining the cause of a crash much easier.
>
> The sensitivity of any data included is minimal as it is the state  
> of the Revolution *engine* that is encoded in these files - this  
> means any state of your Revolution application is completely  
> obfuscated... Unless of course you can divine the functioning of  
> Rev script by the C++ calls that result in getting made in the  
> engine itself.
>
> -- 
>
> Warmest Regards,
>
> Mark.
>
> ***
>
> So it sounds to me like if you want to examine the contents of the  
> log, you could look at it in VB just as Mark does, provided you  
> understand the binary notation.
>
> -- 
> Jacqueline Landman Gay         |     jacque at hyperactivesw.com
> HyperActive Software           |     http://www.hyperactivesw.com



Best regards from Paris,
Eric Chatonet.
----------------------------------------------------------------
Plugins and tutorials for Revolution: http://www.sosmartsoftware.com/
Email: eric.chatonet at sosmartsoftware.com/
----------------------------------------------------------------





More information about the use-livecode mailing list