[Pharo-project] Time Machine
Hi,
From a discussion I had with Kent Beck about Continuous Testing / Autotest (see http://pharocasts.blogspot.com/2010/07/live-testing-with-autotest.html ), I would like to add a kind of Time Machine for Smalltalk code in Autotest.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding. So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change How would you do this ? Where should I look for materials which can help me coding this ? Cheers, Laurent Laffont http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/
sound cool. looks like a research topic on it own.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding.
So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change
How would you do this ? Where should I look for materials which can help me coding this ?
I would use MethodWrapper or the like (may be byetsurgeon when available) and you can also have a look at dynamix (a famix extension in the Phd of orla greevy) to model trace. then you keep track of what changed using a changeset and revert. Stef
- compare execution paths in order to see the consequence of last change
Not easy to get. But yes, this is what the whole world wants :-) Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
did you read delta debugging from andreas Zeller? Stef On Jul 22, 2010, at 11:11 PM, laurent laffont wrote:
Hi,
From a discussion I had with Kent Beck about Continuous Testing / Autotest (see http://pharocasts.blogspot.com/2010/07/live-testing-with-autotest.html ), I would like to add a kind of Time Machine for Smalltalk code in Autotest.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding.
So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change
How would you do this ? Where should I look for materials which can help me coding this ?
Cheers,
Laurent Laffont
http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/ _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Thu, Jul 22, 2010 at 11:34 PM, Stéphane Ducasse < stephane.ducasse@inria.fr> wrote:
did you read delta debugging from andreas Zeller?
No. Where ? Laurent
Stef
On Jul 22, 2010, at 11:11 PM, laurent laffont wrote:
Hi,
From a discussion I had with Kent Beck about Continuous Testing / Autotest (see http://pharocasts.blogspot.com/2010/07/live-testing-with-autotest.html ), I would like to add a kind of Time Machine for Smalltalk code in Autotest.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding.
So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change
How would you do this ? Where should I look for materials which can help me coding this ?
Cheers,
Laurent Laffont
http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/ _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
http://www.st.cs.uni-saarland.de/dd/ you can get some papers but also the slides that go with the book Stef On Jul 22, 2010, at 11:36 PM, laurent laffont wrote:
On Thu, Jul 22, 2010 at 11:34 PM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote: did you read delta debugging from andreas Zeller?
No. Where ?
Laurent
Stef
On Jul 22, 2010, at 11:11 PM, laurent laffont wrote:
Hi,
From a discussion I had with Kent Beck about Continuous Testing / Autotest (see http://pharocasts.blogspot.com/2010/07/live-testing-with-autotest.html ), I would like to add a kind of Time Machine for Smalltalk code in Autotest.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding.
So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change
How would you do this ? Where should I look for materials which can help me coding this ?
Cheers,
Laurent Laffont
http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/ _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
I've ordered "Why programs fail" by Andreas Zeller. It will be my summer holidays book ;) On Thu, Jul 22, 2010 at 11:52 PM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
http://www.st.cs.uni-saarland.de/dd/ you can get some papers but also the slides that go with the book Stef
On Jul 22, 2010, at 11:36 PM, laurent laffont wrote:
On Thu, Jul 22, 2010 at 11:34 PM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote: did you read delta debugging from andreas Zeller?
No. Where ?
Laurent
Stef
On Jul 22, 2010, at 11:11 PM, laurent laffont wrote:
Hi,
From a discussion I had with Kent Beck about Continuous Testing / Autotest (see http://pharocasts.blogspot.com/2010/07/live-testing-with-autotest.html ), I would like to add a kind of Time Machine for Smalltalk code in Autotest.
When a test fails, what I really want to know is why. Sometimes a test fails and I don't know why. So I have to go through a debugging session which distracts me from what I've been coding.
So an idea is to diff backtrace before and after the test fails: - revert to last green - replay the test and save the trace - forward to current test and run (test fails) - compare execution paths in order to see the consequence of last change
How would you do this ? Where should I look for materials which can help me coding this ?
Cheers,
Laurent Laffont
http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/ _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
-- Patrick Barroca
participants (4)
-
Alexandre Bergel -
laurent laffont -
Patrick Barroca -
Stéphane Ducasse