jakarta-bcel-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Peter Schneider" <schne...@lathanda.de>
Subject RE: help with structural analysis
Date Thu, 31 Oct 2002 06:48:39 GMT
Hi
> With multiple paths its quite possible to have multiple frames, so I'd
> happily take the frame with the largest stack.
not really,
valid bytecode must not have different stack size on different pathes!

> At the moment Ive been playing with modifying code in JustIce to allow
> me to examine it interactively, but in the long run I'd like to have a
> batch program I can run over lots of classes.
please read jvm spec chapter 4.8 and 4.9 first. It's really important
for what you are doing.
only about 10 pages. But that knowledge is really important. If you want
to modify the a verifier you can't work without it.
http://java.sun.com/docs/books/vmspec/2nd-edition/html/ClassFile.doc.htm
l#40222

 
> I have three questions:
>  1) has someone done something similar before ?
>  2) should I be using pass3b verifier as a base ?
>  3) if not does anybody have suggestions how to organise it
> 
> And
>  4) what have I missed ?
1) hmm yes, the theory isn't new, i'm not sure if it has been done for
java yet.
2) wouldn't use it. To understand it yes, but not as base for your code.
If you expect that the code is valid, you have less analysis to do.
3) e.g. you can easily find out the stack size for each operation. you
know a lot about jumping behaviour. And if you are happy with sun
compiler, you can expect a lot more.
4) as stated above read that chapter first.

greetings Peter




--
To unsubscribe, e-mail:   <mailto:bcel-user-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:bcel-user-help@jakarta.apache.org>


Mime
View raw message