On Wednesday 05 August 2009 10:08:34 Daniel Dunbar wrote: > On Tue, Aug 4, 2009 at 1:41 PM, Daniel Dunbar<daniel at zuster.org> wrote: > > On Tue, Aug 4, 2009 at 11:56 AM, Vladimir > > > > Kuznetsov<ks.vladimir at gmail.com> wrote: > >> Hello Daniel, > >> > >> Thank you for the explanation. Could you please give a hint how soon the > >> tool will be published ? I am a bit in rush, so I want to decide whether > >> to wait or to implement something more simple for now myself. > > > > Well, if you are in a rush I can check it in this evening; I was > > hoping to clean it up a bit more first but I can do that later. Of > > course, it may or may not work for you, but that is a separate > > problem. :) > > Done. I also wrote the section on the "Testing Coreutils" page on how > to use it to replay tests. It seems to be working for 'echo', at least > on Linux x86-32. I kicked off a coreutils run tonight which I'll > hopefully check the replay on later this week, and also to serve as a > baseline for testing x86-64. > > Let me know if you have any questions regarding the klee-replay tool, > and if it works for you.
Thank you very much. The tool works well for me so far. > > - Daniel > > >> -- > >> With the best regards, > >> Vladimir Kuznetsov > >> > >> On Tuesday 04 August 2009 18:54:18 Daniel Dunbar wrote: > >>> On Tue, Aug 4, 2009 at 3:25 AM, Vladimir > >>> Kuznetsov<ks.vladimir at gmail.com> > >> > >> wrote: > >>> > Hello, > >>> > > >>> > What is the easiest way to replay tests with symbolic arguments and > >>> > symbolic files ? It seems that libkleeRuntest.so is not handling > >>> > them. One way would be to parse .ktest file and manually invoke a > >>> > command with the arguments from the file. Is there a better solution > >>> > ? > >>> > >>> Hi Vladimir, > >>> > >>> We have a tool which does basically that, it creates the symbolic > >>> files, sets up pipes, etc and then runs the program. It also allows us > >>> to test system call failures by using ptrace to control the > >>> subprocess, which is convenient but adds a lot of complexity. > >>> > >>> I started working on cleanup up the code and making it a little easier > >>> to use, but it isn't quite ready yet. I'm hoping to get a basic > >>> version in the repo fairly soon. > >>> > >>> - Daniel > >>> > >>> > -- > >>> > With the best regards, > >>> > Vladimir Kuznetsov > >>> > _______________________________________________ > >>> > klee-dev mailing list > >>> > klee-dev at keeda.stanford.edu > >>> > http://keeda.Stanford.EDU/mailman/listinfo/klee-dev -- With the best regards, Vladimir Kuznetsov
