ForLoopIncrementInUpdate should refrain from moving increment to the update part of the for in case the variable is incremented multiple times

What problem are you trying to solve?

I want to keep the code reasonable, and ForLoopIncrementInUpdate produces suboptimal results.

Describe the solution you'd like

PreparedStatement ps
...
      // Multi values(),(),() case
      for (int i = 0; i < p1nrows; ) {
        for (int k = 0, pos = 1; k < p2multi; k++, i++) {
          ps.setInt(pos, i);
          pos++;
          ps.setString(pos, strings[i]);
          pos++;
          ps.setInt(pos, i);
          pos++;
        }
        ps.addBatch();
      }

Unfortunately, ForLoopIncrementInUpdate moves the last increment of pos to the for which makes the code awkward:

      for (int i = 0; i < p1nrows; ) {
        for (int k = 0, pos = 1; k < p2multi; k++, pos++, i++) { // <-- pos++ is weird here
          ps.setInt(pos, i);
          pos++;
          ps.setString(pos, strings[i]);
          pos++;
          ps.setInt(pos, i);
        }
        ps.addBatch();
      }

I think it would be reasonable if ForLoopIncrementInUpdate does not touch variables that are incremented several times in the loop body.

Have you considered any alternatives or workarounds?

THere's an option to ignore and accept the change.

Additional context

https://github.com/pgjdbc/pgjdbc/pull/2979

Start by locating the ForLoopIncrementInUpdate recipe and inspect how it identifies increments moved into a for-loop update clause. Use the provided nested-loop example as the reproduction case. Done means variables incremented multiple times in the loop body are left in the body, while eligible increments still behave as before; add or run the relevant recipe tests once located.