explain ConfigurationOptions

[yikengboy]

Can you explain what impact adjusting the bu_minsim threshold will have on the output of the editing script?
I debug source code, I found the place where the threshold is used. It seems that the larger the value, the less mapping there is in the mapping.

I tested it and found that the larger the threshold, the more types the editing script returns. Why is this?

[jrfaller]

Hi! I advise you to look into our papers to better understand these thresholds.

Note that there is a new undocumented auto generator that setup this threshold for you automatically at the expense of runtime!

Jean-Rémy Falleri, Floréal Morandat, Xavier Blanc, Matias Martinez, Martin Monperrus:
Fine-grained and accurate source code differencing. ASE 2014: 313-324

Matias Martinez, Jean-Rémy Falleri, Martin Monperrus:
Hyperparameter Optimization for AST Differencing. IEEE Trans. Software Eng. 49(10): 4814-4828 (2023)

Jean-Rémy Falleri, Matias Martinez:
Fine-grained, accurate and scalable source differencing. ICSE 2024: 231:1-231:12

[yikengboy]

OK, I will study your paper. by the way, how to use this new undocumented auto generator.

[yikengboy]

I browsed the implementation of Match and found SimpleGumtreeAutoMt matcher,
this matcher using different thresholds, the final output returns the smallest capacity edit script, which is exactly what I want.
Is this what automatic generator you replied to above?

[jrfaller]

Yes it's exactly that!

[yikengboy]

okay. Thanks a lot. 😁
However, I feel that these thresholds should be dynamically generated first, and then the mapping calculated. Maybe this will make the performance faster.