Re: What's cooking in git.git (Feb 2018, #02; Tue, 13)

Derrick Stolee [off-list ref] writes:

There have been a few "What's cooking" emails since I submitted v1 of
"Serialized Git Commit Graph" and it has not appeared with a tracking
branch. Is this a mistake, or is it something about the state of the
review?

The latter.

Once I pick up a topic and have it in 'pu', I'd be committing to
carrying it and keeping it up-to-date, while dealing with possible
conflicts with other topics.  As I do not have infinite bandwidth, I
try not to chase targets that are still moving too rapidly, which in
turn means that a hot topic everybody is excited by its goal will
take more rerolls than other topics before hitting 'pu', because it
gets more good suggestions and it takes time for its patches to stop
morphing a lot.

The discussion in the last and current rounds gave me an impression
that some stuff (e.g. "graph-head") are still likely to change quite
a lot during the review-response cycle.  Is everybody happy with the
latest set of patches or are there issues raised already in the
review that are better addressed before we start making it interact
with other topics in flight?

Thanks for pinging.

On 2/14/2018 12:23 PM, Junio C Hamano wrote:

Derrick Stolee [off-list ref] writes:

quoted

There have been a few "What's cooking" emails since I submitted v1 of
"Serialized Git Commit Graph" and it has not appeared with a tracking
branch. Is this a mistake, or is it something about the state of the
review?

The latter.

Once I pick up a topic and have it in 'pu', I'd be committing to
carrying it and keeping it up-to-date, while dealing with possible
conflicts with other topics.  As I do not have infinite bandwidth, I
try not to chase targets that are still moving too rapidly, which in
turn means that a hot topic everybody is excited by its goal will
take more rerolls than other topics before hitting 'pu', because it
gets more good suggestions and it takes time for its patches to stop
morphing a lot.

Thanks for clarifying. That makes sense.

The discussion in the last and current rounds gave me an impression
that some stuff (e.g. "graph-head") are still likely to change quite
a lot during the review-response cycle.  Is everybody happy with the
latest set of patches or are there issues raised already in the
review that are better addressed before we start making it interact
with other topics in flight?

To avoid causing a tangent in this thread, I'll send a message on the v3 
thread summarizing what I plan to do for v4 and ask for consensus on 
that approach before I do.

Thanks,
-Stolee