Re: timed contains buffer overflows, and more
:I fully agree with you. However it seemed more applicable to put it all in
:one large patch, because that takes relatively less time to check than
:several seperate patches. I will do this in the future, but my main motive
No, it definitely isn't correct to put multi-categorical changes into
one patch, because other projects will often want to port certain aspects
of our work into their codebase and mixing multi-categorical changes
into a single patch will result in those projects having to spend
a long time picking out the pieces they want (such as strcpy->strlcpy)
from the pieces they don't (all the random non-functional cleanups).
So, for example, I think it would be ok to mix (void) removal with,
say, declaration/assignment separation, but it would not be ok to mix
either of those with functional changes like strcpy->strlcpy. You could
mix strcpy->strlcpy and sprintf->snprintf changes together into one
patch but you should not mix that with, unrelated cleanups.
As it is I am going to have to separate out the timed work into two
separate commits. I would prefer not to have to do that again because
it is going to take two hours of my time to fix just timed and it doesn't
really help the project out if the patches you submit require a committer
to spend as long or longer reviewing and/or separating them.
-Matt
Matthew Dillon
<dillon@backplane.com>
:is that no patches are lost (like has happened in the future with my
:patches - rmdir is still not committed) and that they are committed as soon
:as possible.. I think you'll catch my drift.
:
:Thanks for the pointers anyway, will do in the future. And thanks for taking
:the time to review my patches.
:
:Rgds,
:Douwe
討論串 (同標題文章)
完整討論串 (本文為第 6 之 7 篇)：