011. July 9 to July 13

Planned tasks for this week

• task 1: continue adding discussed edits to paper ⌛️
• task 2: discuss which graphs work best ✔️
• task 3: figuring out if Gumtree can give us qualified api names ⌛️
• task 4: read the methodology of the related work ✔️

Progress

task 1

• in the process of adding edits from feedback and improving paper in general
• added a table of APIs that we've seen, as discussed
• done adding suggestions from last week (as best as I could)
• currently working on suggestions from this thursday

task 2

• Discussed at thurs. meeting
  • removing outliers
  • separating APIs by tag
  • violin plots by tag
  • reverse stacked bars by star rating
  • inferences we can make from the data (i.e. x by y vs. y by x)
  • fix the messed up UX on the graphs

task 3

• emailed Dr. Falleri from GumTreeDiff - forwarded both responses (July 12 7:29 am & 12:38 pm)
• Have tried:
  • computing the Actions between two files where the change was String -> StringBuilder: The output includes "UPD 43@@String from String to StringBuilder", but not the full API name
  • using toTreeString on an ITree: prints a representation of the whole tree, but doesn't show the full API names either
  • getting a JDTVisitor, which has getFullyQualifiedName, integrated with the ITree
• Trying:
  • getLabel() returns the node's original content (e.x., "43@@String") for an ITree. Maybe we can just fork the repo and refine the ITree getLabel method to give us the full label?
  • What Dr. Falleri suggests in second email response + Emerson's reply
    • was able to create AST for the test file, but still working on getting names; having some trouble with the resolved bindings
  • (if nothing works try another tool)

task 4

• have read the paper from MSR 2018
  • currently making a comparison summary with this paper and others we've looked at during this summer

Other

• wrote some scripts to count data properties

Open Problems

• new issues opened this week: NONE
• Each GumTreeDiff node has a type (int) like "43@@ ...", but I'm not entirely sure what it means? It doesn't seem to be a line number or a node number from looking at the output of toTreeString.

Things we discussed/agreed on

• automating the data counting process
• also see task 2 and 4

Next steps

• continue working with GumTreeDiff given new info
• separate api's by tag
• fix graphs