Dear Brad,
in message 17415.38634.967662.176108@waldo.lisle.iphase.com you wrote:
I see there has been some promising commit activity in your U-Boot git tree over the past couple of weeks, tho it is not always clear to me what drives the choice of patches being applied.
There are several factors:
* customers who pay for a port obviously receive highest priority
* all remaining patches are normally processed sequentially, as time permits
* However, there are a couple of non-trivial issues or problem areas where many patches have been submitted that affect the same area of code; often an additional complication is that I have only limited or no access to hardware that can be used to test the suggested fixes. Such areas are NAND support and the CFI driver.
For NAND support, I tried how it works to use a testing branch. This was somewhat frustrating, as feedback from the board maintainers was epsilon.
For CFI suport, Stefan Roese volunteered to integrate all submitted patches. Stefan is a precious, but unfortunately limited resource.
Can you give us some statistics on the state of your patch queue and perhaps an idea about when you might get to the bulk of the patches submitted during the last quarter of last year (and earlier)?
I have it on my list every day (and night) and weekend, but always other, higher prirority things need to be done first.
If you don't reject a patch immediately on submission is it likely it will be committed some day?
Probably. In case of complexer patches I often don't have the time to perform thorough initial checking, so it may happen that I chose to reject such apatch later, when I try to add it. But ideally all submitted patches are in my queue and will be processed one day.
Should I wait to submit patches which depend on patches already submitted by others or just assume the previous patches will be committed eventually?
You can assume that previous patches will be applied, but you do yourself a favour if you mention any such prerequisites in your own patch submission.
Best regards,
Wolfgang Denk