[U-Boot] [PATCH] Fix spelling of "occurred".

[Simon Glass]

Hi Scott,

On 2 May 2016 at 13:03, Scott Wood <oss at buserror.net> wrote:
> On Mon, 2016-05-02 at 12:57 -0600, Simon Glass wrote:
>> Hi Scott,
>>
>> On 1 May 2016 at 17:34, Scott Wood <oss at buserror.net> wrote:
>> > On Sun, 2016-05-01 at 12:55 -0600, Simon Glass wrote:
>> > > Hi,
>> > >
>> > > On 30 April 2016 at 20:18, Vagrant Cascadian <vagrant at debian.org> wrote:
>> > >
>> > > Please can you add a commit message?
>> >
>> > I don't understand these "empty/missing commit message" remarks when
>> > there's a
>> > one-line changelog (in the subject).  Do you seriously want the same line
>> > repeated twice in the git commit, just so something shows up in the body
>> > of
>> > the e-mail?  It's one thing if the commit warrants more than a single line
>> > (though it's still not accurate to say that the changelog is completely
>> > absent), but a spelling fix is about as trivial as it gets...
>> > -Scott
>> >
>>
>> It only takes a few seconds to add a commit message and I think it is
>> good practice.
>>
>> But if you want to allow commits with no message (other than
>> merge/release tag), then we should document it here:
>> http://www.denx.de/wiki/U-Boot/Patches
>
> There is a commit message.  It is 'Fix spelling of "occurred"'.
>
> And that wiki link explicitly says, "Put a detailed description after the
> summary and blank line. If the summary line is sufficient to describe the
> change, you can omit the blank line and detailed description."

OK I made a little update to make it more limited:

"Put a detailed description after the summary and blank line. If the
summary line is sufficient to describe the change (e.g. it is a
trivial spelling correction or whitespace update), you can omit the
blank line and detailed description."

Does that seem reasonable?

We should avoid submitting new drivers and forgetting a commit
message. Also some fixes are trivial (e.g. adding 1 to a result) but
the reason for them needs to be explained. I'd rather have a hard and
fast rule but I agree in some cases it doesn't make sense and it needs
to be flexible.

Sorry if I'm being picky on this. I've spent a *lot* of time over the
past few years digging into code and finding changes that are not
self-explanatory. That's why I am keen on people adding comments to
header files for API functions and struct members

 I think it helps for people to answer the question 'why should this
code be submitted?' when writing a commit message. For the record. in
this case, I would have added a message something like 'Occurred is
spelled incorrectly in a number of places. Fix these up to provide
consistency'.

Regards,
Simon