40
votes

I want to use clang-format to align my comments, but nothing else.

The option for that is: AlignTrailingComments (bool).

But when I run the following:

clang-format-3.6 -i -style='{AlignTrailingComments: true}' <file>

It performs all kinds of other formatting options that I suppose have a default when unspecified.

How can I execute just one clang formatting rule on the codebase?

Having all of these defaults make it difficult to see the full effect that a single formatting option has on the code. I have to parse through the diff of all these other changes and decide if it was the option I specified that actually did it.


I noticed that there is a DisableFormat option, but no matter how I use it, it stops any formatting from happening at all.

clang-format-3.6 -i -style='{AlignTrailingComments: true, DisableFormat: true}'

clang-format-3.6 -i -style='{DisableFormat: true, AlignTrailingComments: true}'

Both cause clang-format to not make any code anywhere.

2
i think you should specify BasedOnStylebolov
you should try it as a plugin (e.g. for Vim , Visual-Studio). It is much handy to use this way.bolov
@bolov As a plugin, will it still apply more than one formatting rule? There are many conflicting formatting styles in our codebase, and I don't want to clutter a commit with all of these additional changes that effect other people's portions of code. I just want to apply a single style change so that other people can consume the change and handle merge conflicts more easily. It's not practical to subscribe to an entire code style using BasedOnStyle right now.Trevor Hickey
@bolov Because I only want to apply one rule change at at a time. I need a default configuration where clang-format won't make any changes at all. Then I can tweak one at a time, and see a diff of only that format option.Trevor Hickey
Blegh, came here for this functionality too. Sad it doesn't exist.Barry

2 Answers

19
votes

I think clang format was simply not designed for this. Rules are not things it applies incrementally, the program is instead built around, parsing your entire program and forgetting (most of) the old whitespace, and generating new whitespace based on the rules you select.

You can see some overview of the architecture here: http://www.llvm.org/devmtg/2013-04/jasper-slides.pdf

First it runs clang lexer and parser, then it divides groups of tokens into "unwrapped lines" which are "tokens we would like to put together on a single line if there was no column limit". Then a layouter determines the formatting of each unwrapped line based on the various constraints and optimizing for the various penalties.

So, I don't think "one clang-format action" is actually a thing, the design looks pretty much monolithic to me.

3
votes

clang-format does not have a mechanism to do this. The palliative I use is to configure what I want changed in two different ways, I run the tool with both configurations and diff them, that tells me the changes that affect what I want, but in general this requires at least some manual work.

clang-format is a good idea with a poor implementation. Some of the many deficiencies it has are that there is no option to have it indicate the reason for a change, the XML output format is not consumable by any other popular tool...