Martin Thomson writes: > I would like to think that adding (or removing) braces from block statements > should be acceptable.
I would argue that braces should not be added with automation. When debugging code, it is important to understand the intent of the author. Adding braces at parser-determined points disguises the intent. Actually, the same argument would prefer that code should not be automatically re-indented to parser-determined tabs. > The tool that I posted a link to earlier has ways to maintain blame across > both of the above sorts of changes. I get the impression that blame-bridge's goal is to attribute lines to the correct person, but what we want is to identify the changeset that made the code as it is now. I assume that can be done if we accept another level of indirection. This won't help in-progress patches, anyway. > I disagree. I think that the bigger costs come from changing naming > conventions (_memberVariable to memberVariable_ to mMemberVariable), because > that can’t be done a file at a time. We are talking about different things here. Yes, it will be harder to automatically fix up patches for variable renaming than for whitespace changes, but at least the compiler will usually detect problems with different variable names. _memberVariable to memberVariable_ to mMemberVariable is another situation where the particular choice does not matter. All are readable, so the only cost is when moving code from a file that uses memberVariable_ to one that uses mMemberVariable. That is not something we do too often. _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
