Hello,
Thank you for your answer, it helped me find the differences between the critical pairs. So it seems that their critical elements are different. So do I understand right: for each rule that there are conflicts it will
return critical pairs and they store a particular node in the critical elements that would cause the conflict?
So let's say I have my input model and I generated the matches for the rules and I want to check if these two particular matches have conflicts. Then I'd just have to look into which ones are the critical elements and
check if those are the same for my match? Do I understand it correctly?
Thank you,
From: henshin-dev-bounces@xxxxxxxxxxx <henshin-dev-bounces@xxxxxxxxxxx> on behalf of Daniel Strüber <strueber@xxxxxxxxxxxxxxxxxxxxxxxxx>
Sent: 10 March 2019 23:08:55
To: henshin-dev@xxxxxxxxxxx
Subject: Re: [henshin-dev] Henshin CPA
Hi Kinga,
On 11.03.2019 00:01, Bojarczuk, Kinga wrote:
Hello,
Thank you for your answer. I have seen the wiki page before. So perhaps I will just list my questions instead and it'll be easier:
1) I'm running runConflictAnalysis on a set of rules where only a pair of rules called 'assemble' and 'assemble' should return conflicts. What I don't quite understand is why the runConflictAnalysis method returns a
list of 8 pairs where in each pair it is the same 'assemble' rule with delete-use-conflict. I don't understand why it doesn't just return it once?
Have you actually inspected the results very closely to be absolutely sure that they're the same? Often there are multiple ways to construct a conflict which are different only in a very subtle way (that comes from the underlying theory of critical pairs).
With the recent work I mentioned, we will provide more easy-to-understand results in which such very similar results are shown as one result.
2) What I'd like to be able to do is using the current state and knowing which rules have matches, be able to run the analysis and check if two particular matches for a rule have conflicts. Am I able to do that somehow?
"knowing which rules have matches" - matches to what exactly? Is there an input model in your case? CPA is a static analysis that doesn't take an input model, but just a pair of rules. It constructs minimal input models for which a conflict of the rules would
arise.
If an input model is given, you need another kind of analysis, one that is potentially easy to implement (we have code for finding all matches, and you would need to check if the provided matches overlap, by iterating over them).
Best regards,
Daneil
Hi Kinga,
since you don't mention it explicitly, I wonder if you had a look at the Wiki page yet:
https://wiki.eclipse.org/Henshin/Critical_Pair_Analysis
I think that it address your questions. But, if the explanations are not clear enough, feel free to follow up.
Best regards,
Daniel
On 08.03.2019 16:35, Bojarczuk, Kinga wrote:
Hello,
I'm working on a final year project with Steffen. I need the henshin cpa package for my project and I need to use it programmatically. I cannot seem to find the docs / info about it online. I've found some but only some
very basic stuff. I already have code which uses the package and does the analysis but for my project I'd like to be able to understand exactly how the runConflictAnalysis and runDependencyAnalysis methods work and what they return as I'm a bit confused about
that, especially the 'matches' that the CriticalPairs store after running the analysis. Could you please point me to some resources ?
Thank you,
_______________________________________________
henshin-dev mailing list
henshin-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/henshin-dev
_______________________________________________
henshin-dev mailing list
henshin-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/henshin-dev
|
