[Maria-discuss] community contribution measurement page
Folks,
I've prepared a community contribution measurement page:
http://measure-mariadb.ozlabs.org/repo/mariadb/server-outside-org.html
This is largely the code base of https://github.com/MeasureOSS/Measure
known errors: * I've yet to create the organisation aspects of this site. * some top page may be distorted by non-mariadb respositories
Insights and suggested improvements welcome (especially with code) :-)
Hi, Daniel!
Wow, this is very informative. Thanks a lot!
It's nice to see how "First PR response" time got improved (in 2015?) after Svoj started handling PRs.
Recent increase in PRs can, perhaps, be partially attributed to Tempesta and Codership? And, perhaps, partially because we're getting more PRs now and need to change internal procedures to be able to cope with them?
I don't understand the "PRs open and closed this month" chart, number of "opened PRs" seems to be always less than the number of "closed PRs", how is this possible?
We seem to be accepting about 60% of all PRs? It looks like since 2016 we accept a bit more than a half. I think it's a pretty good ratio, meaning we must have rather smart contributors who create really good pull requests :)
On Jul 22, Daniel Black wrote:
Folks,
I've prepared a community contribution measurement page:
http://measure-mariadb.ozlabs.org/repo/mariadb/server-outside-org.html
This is largely the code base of https://github.com/MeasureOSS/Measure
known errors: * I've yet to create the organisation aspects of this site. * some top page may be distorted by non-mariadb respositories
Insights and suggested improvements welcome (especially with code) :-)
Regards, Sergei Chief Architect MariaDB and security@mariadb.org
Hi Daniel!
Some pr's get merged by a separate commit and the issue itself gets closed, not merged on github. Do your graphs catch that use case?
Vicentiu
On Sun, 22 Jul 2018, 14:30 Sergei Golubchik, serg@mariadb.org wrote:
Hi, Daniel!
Wow, this is very informative. Thanks a lot!
It's nice to see how "First PR response" time got improved (in 2015?) after Svoj started handling PRs.
Recent increase in PRs can, perhaps, be partially attributed to Tempesta and Codership? And, perhaps, partially because we're getting more PRs now and need to change internal procedures to be able to cope with them?
I don't understand the "PRs open and closed this month" chart, number of "opened PRs" seems to be always less than the number of "closed PRs", how is this possible?
We seem to be accepting about 60% of all PRs? It looks like since 2016 we accept a bit more than a half. I think it's a pretty good ratio, meaning we must have rather smart contributors who create really good pull requests :)
On Jul 22, Daniel Black wrote:
Folks,
I've prepared a community contribution measurement page:
http://measure-mariadb.ozlabs.org/repo/mariadb/server-outside-org.html
This is largely the code base of https://github.com/MeasureOSS/Measure
known errors: * I've yet to create the organisation aspects of this site. * some top page may be distorted by non-mariadb respositories
Insights and suggested improvements welcome (especially with code) :-)
Regards, Sergei Chief Architect MariaDB and security@mariadb.org
_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
On Sun, 22 Jul 2018 14:32:42 +0300 Vicențiu Ciorbaru cvicentiu@gmail.com wrote:
Hi Daniel!
Some pr's get merged by a separate commit and the issue itself gets closed, not merged on github. Do your graphs catch that use case?
Most likely not - the data is largely based on what github exposes in their API. Perhaps a less overloaded word than 'rejected' or a note that closed isn't necessarily abandoned/rejected.
Commit data is there so is there an almost reliable pattern for this? Who closes it perhaps?
I assume the cases here are a rebase of a form where the commit IDs don't match (because I think github detects that as a merge).
Vicentiu
On Sun, 22 Jul 2018 13:29:07 +0200 Sergei Golubchik serg@mariadb.org wrote:
Hi, Daniel!
Wow, this is very informative. Thanks a lot!
You're welcome.
It's nice to see how "First PR response" time got improved (in 2015?) after Svoj started handling PRs.
True. Was that a foundation goal at the time?
Recent increase in PRs can, perhaps, be partially attributed to Tempesta and Codership?
Maybe, if i get the Team/Organisations going this will be a bit more obvious.
And, perhaps, partially because we're getting more PRs now and need to change internal procedures to be able to cope with them?
In addition to growth, there is a 95th percentile of 25 days for open prs in contrast with the median of 160 days shows a number aren't being reviewed/rejected/merged.
If you can come up with some chart/goal that aids the internal procedures then I can either try to develop or publish it.
Possibly something around the time in a github PR review state, or releases passed for an approved PR (even though these aren't currently scraped).
I don't understand the "PRs open and closed this month" chart, number of "opened PRs" seems to be always less than the number of "closed PRs", how is this possible?
Its a stacked graphs but you're right, something's odd there, I'll take a look. https://github.com/MeasureOSS/Measure/blob/master/widgets/repo/43_prsStackCh...
We seem to be accepting about 60% of all PRs? It looks like since 2016 we accept a bit more than a half. I think it's a pretty good ratio, meaning we must have rather smart contributors who create really good pull requests :)
Given the amount of feedback and contributor behaviour to create a new pr or git push, or merged/rebased in background (as Vicențiu suggested), I think 60% is something to be proud of.
As a measure subject to varying effects, its relative performance probably would need to be very careful about inferences made.
On Jul 22, Daniel Black wrote:
Folks,
I've prepared a community contribution measurement page:
http://measure-mariadb.ozlabs.org/repo/mariadb/server-outside-org.html
This is largely the code base of https://github.com/MeasureOSS/Measure
known errors: * I've yet to create the organisation aspects of this site. * some top page may be distorted by non-mariadb respositories
Insights and suggested improvements welcome (especially with code) :-)
Regards, Sergei Chief Architect MariaDB and security@mariadb.org
participants (3)
-
Daniel Black -
Sergei Golubchik -
Vicențiu Ciorbaru