Hi Tim, On 14.01.2026 19:26, Tim Perry wrote: > Is the problem RTC? Or has better become the enemy of good enough in the > reviews? I know I'm walking a fine line, but sometimes the best way to > handle issues discovered in a code review is to open a separate ticket for > that work. For me, the request to change the handling of paths as strings > that was mentioned earlier falls into that category.... Sure, the person > trying to change the code could clean up issues in the surrounding code, > but must they? Seems like a high hurdle.
Regarding PR 3855 [1], I had several discussions with Gary about it (for example, [2]). My final proposal was not to block the feature itself, but to ask for a small generalization: adding the same `getPath` method to `RandomAccessFileManager` and `MemoryMappedFileManager`, so that all file appender implementations retain feature parity, not just the one Gary happens to use most often. Looking back, I think there were missteps on both sides in how this played out: - I initially misunderstood the context and assumed the feature was directly required by Gary’s employer, which led me to ask whether a broader change, more useful to the community, could be contributed instead. - Gary, on the other hand, did not escalate the disagreement to this mailing list and instead disengaged from the discussion. Ultimately, we’re a consensus-seeking (or at least majority-seeking) community. Apache Logging Services has 47 committers whose opinions all carry weight, and decisions aren’t made by any single reviewer. In this case, the PR received reviews from Volkan and myself, but no others. I don’t think it’s reasonable to blame us for the lack of other reviews. Piotr [1] https://github.com/apache/logging-log4j2/pull/3855 [2] https://lists.apache.org/thread/8xpg47q7wqvvhw9myqh6fzc8lmotgsz6
