Re: [PATCH 00/16] Fixes; Document remaining stdint.h types

[Alejandro Colomar <colomar.6.4.3@gmail.com>]

Okay then :)

Thanks,

Alex

On 2020-10-01 13:50, Michael Kerrisk (man-pages) wrote:
> Hi Alex,
> 
> On 10/1/20 1:41 PM, Alejandro Colomar wrote:
>> Hi Michael,
>>
>> I did it this way because then you have a clearly ordered list
>> of the commits, and in which order they go,
>> so I thought it might be easier for you (creating less conflicts).
> 
> Yes, I understand the rationale. But when I get a series of
> loosely related patches in a series of 20, and multiple
> conversations start about independent topics, I'm finding
> it quite some effort to keep track.
> 
>> And also, I can hold any more recent patches, such as __int128,
>> for when you finish applying the previous set, so I fix the
>> conflicts before you ever see them.
>>
>> Don't you think?
>>
>> I don't mind fixing for example patch 5,
>> and then rebasing the rest (and also the patches I didn't send yet),
>> and resending them as an answer to v1 00/16.
>>
>> But if you still prefer smaller sets, I'll send you smaller sets.
> 
> I do prefer smaller sets. And yes, occasionally things may
> go wrong in terms of patch conflicts, but I think that may be
> a smaller than the problem I note above.
> 
>> It's just that these patches are usually very dependent of the
>> previous ones, and therefore prone to conflicts if you
>> don't apply them in the same exact order.
>>
>> Your thoughts?
> 
> As you can see, there's no perfect solution here. In such
> situations what I try to do (where possible) is order the
> patches from least contentious to most contentious.
> That way, the patches that are almost certainly going to
> be applied are loaded at the front and the chance of having
> to rebasing later patches in a series is lower.
> 
> Thanks,
> 
> Michael
> 
> 
> 
>> On 2020-10-01 13:32, Michael Kerrisk (man-pages) wrote:
>>> Hi Alex,
>>>
>>> On 10/1/20 12:15 PM, Alejandro Colomar wrote:
>>>> Hi Michael,
>>>>
>>>> Here are a few fixes (including one removing .br),
>>>> and then the remaining stdint types.
>>>
>>> These very long patch series are a bit overwhelming for me.
>>> I'd have preferred a few smaller patch series. For example,
>>> I think I would have preferred 3 series like this:
>>>
>>> 1-4
>>> 5-12
>>> 13-16
>>>
>>> One reason is that the multiple parallel reply threads that
>>> sometimes occur can sometimes be rather difficult to track.
>>> (Your patches have started some quite useful conversations!)
>>>
>>> For example, I suspect Jonathan's comments may trigger changes
>>> for patches 5-12.
>>>
>>> For now, I'm applying 1-4 and 13-16. It looks like some reworking is
>>> going to be needed for the others. When you do resubmit them, please
>>> start a new thread (rather than replying into this thread).
>>>
>>> Thanks,
>>>
>>> Michael