5 ms·
A recent rant from Linus Torvalds on commenting style: https://lkml.org/lkml/2016/7/8/625 https://lkml.org/lkml/2016/7/8/625
by 0xmohit 10y ago
A recent rant from Linus Torvalds on commenting style: https://lkml.org/lkml/2016/7/8/625 https://lkml.org/lkml/2016/7/8/625
- arethuza 10y agoWhat's so bad about that? It actually seems like a pretty sensible and well argued bit of feedback - with some slightly colorful language but nothing that seems that extreme.
- skoczymroczny 10y agoHe said "brain-damaged" which probably offended some sensitive folks.
- soundwave106 10y agoIt's honestly rather long winded for something the compiler isn't going to give two hoots about. :) When it comes to syntax, I personally think that the best policy is to pick one style, and stick with it. I have come to distrust people with passionate rants like this about syntax because, in my experience, in a few years there probably will be another syntax that falls in fashion, and all of a sudden this New Syntax that now is The New One True Way that everyone gets passionate about, and you have to be high to actually like Old Syntax or something. What can happen in long term projects is syntax mishmash. Not terribly a big deal with comment formatting, but annoying for other things. Wonder if that variable you are trying to remember is camelCase, underscore_notation, or sHungarianNotation? Is the table you are joining to using Id as a key, TableId as a key, or Table_Id as a key? That's more annoying than any of the problems any syntax has. So Linus's rant only would be useful to me if his opinion on comment formatting has been consistent over the years. :)
- 0xmohit 10y agoNothing bad. It would have been good if the kernel folks could have written or made use of an existing tool (e.g. GNU indent) that would ensure that the style is consistent. Inconsistent style hurts the eyes.