How do I use patchew to be a good block-devel citizen? (was: Re: [PATCH] iotests/026: Move v3-exclusive test to new file)

[John Snow]

On 3/23/20 8:23 AM, Max Reitz wrote:
> On 12.03.20 23:19, John Snow wrote:
>>
>>
>> On 3/11/20 10:07 AM, Max Reitz wrote:
>>> data_file does not work with v2, and we probably want 026 to keep
>>> working for v2 images.  Thus, open a new file for v3-exclusive error
>>> path test cases.
>>>
>>> Fixes: 81311255f217859413c94f2cd9cebf2684bbda94
>>>        (“iotests/026: Test EIO on allocation in a data-file”)
>>> Signed-off-by: Max Reitz <address@hidden>
>>
>> Let me start this reply with something good, or at least something
>> that's not bad. It's value neutral at worst.
>>
>> Reviewed-by: John Snow <address@hidden>
>> Tested-by: John Snow <address@hidden>
> 
> Thanks. :)
> 
>> Now, let's get cracking on some prime nonsense.
>>
>> I assume this patch is still 'pending'. Here's a complete tangent
>> unrelated to your patch in every single way:
> 
> Reasonable, but it’s a bit of a shame you bury it here.  I feel like
> this makes it unlikely to reach the people you want to reach.
> 

I'm not sure if I was trying to reach anyone, exactly, but maybe I
should have been. Let's recirculate this to a wider audience.

So, directly below, find the $subject question:

>> What's the best way to use patchew to see series that are "pending" in
>> some way? I'd like to:
>>
>> - Search only the block list (to:address@hidden. I assume this
>> catches CCs too.)
>> - Exclude series that are merged (-is:merged)
>> - Exclude obsoleted series (-is:obsolete)
>>
>> This gets a bit closer to things that are interesting in some way --
>> give or take some fuzziness with patchew's detection of "merged" or
>> "obsoleted" sometimes.
>>
>> - Exclude pull requests. (-is:pull seems broken, actually.)
>> - Exclude reviewed series (-is:reviewed -- what does patchew consider
>> 'reviewed'? does this mean fully reviewed, or any reviews?)
>>
>> This gives me something a bit more useful.
>>
>> - Exclude 'expired' series. I use 30 days as a mental model for this. It
>> might be nice to formalize this and mark patches that received no
>> replies and didn't detect any other state change as "expired" and send
>> an autoreply from the bot.
>>
>> (I.e., patches that are complete, applied, passed CI, were not
>> obsoleted, did not appear to be merged, and received no replies from
>> anyone except the patch author)
>>
>>
>> ("Hi, this patch received no replies from anyone except the author (you)
>> for 30 days. The series is being dropped from the pending queue and is
>> being marked expired. If the patches are still important, please rebase
>> them and re-send to the list.
>>
>> Please use scripts/get_maintainers.pl to identify candidate maintainers
>> and reviewers and make sure they are CC'd.
>>
>> This series appears to touch files owned by the following maintainers:
>> - Blah
>> - Etc
>> - And so on
>>
>> For more information on the contribution process, please visit:
>> <wiki links to contribution guides, etc>")
>>
>> We don't have anything like that, so age:<30d suffices. Alright, this
>> list is starting to look *pretty* decent.
>>
>> project:QEMU to:address@hidden not:obsolete not:merged
>> -is:reviewed age:<30d
>>
>> Lastly, maybe we can exclude series that don't have replies yet.
> 
> Maybe Patchew should ping these after 14 days or so.
> 
> That’s about the only thing I can contribute, because I don’t really use
> Patchew myself...  I keep patches in a subfolder of my inbox on unread;
> and I keep my own pending series in a separate folder.  (Before I did
> that, I was much more prone to forgetting my own patches rather than
> someone else’s.)
> 
> Max
> 
>> It's
>> not clear to patchew which replies are:
>>
>> - Unrelated comments, like this one here
>> - Requests for a change
>> - A question for the submitter
>> - A softly-worded N-A-C-K
>>
>> and without a concept of designated reviewer, perhaps lack of replies is
>> good evidence that the series is untouched and needs someone to 'pick it
>> up'; (-has:replies)
>>
>> https://patchew.org/search?q=project%3AQEMU+to%3Aqemu-block%40nongnu.org+not%3Aobsolete+not%3Amerged+-is%3Areviewed+age%3A%3C30d+-has%3Areplies
>>
>> Alright, that's pretty good, actually.
>>
>> OK, yes, this patch still needs love as far as patchew understands.
>>

[Snip: no longer relevant to the topic.]