Including Original Author in git commits.
Hi all, Right now we are committing with commitrer name as author and adding contributor name in commit message. This way of committing will not reflect in contributors git profile. I believe current way of crediting.. is based on SVN days.. Since now all apache repos are following git.. we can use git functionalities (already some are using PRs) and use original author name in commits and use signed of message for committers. Any thoughts? -Vinay
Re: Including Original Author in git commits.
Hi Vinay, I'm already doing this if I can get the original author name and the email address in some way. If the patch is created by git format-patch command, smart-apply-patch --committer option can do this automatically. Thanks, Akira On Thu, Feb 14, 2019 at 3:27 PM Vinayakumar B wrote: > > Hi all, > > Right now we are committing with commitrer name as author and adding > contributor name in commit message. This way of committing will not reflect > in contributors git profile. > > I believe current way of crediting.. is based on SVN days.. > > Since now all apache repos are following git.. we can use git > functionalities (already some are using PRs) and use original author name > in commits and use signed of message for committers. > > > Any thoughts? > > -Vinay - To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org For additional commands, e-mail: common-dev-h...@hadoop.apache.org
Re: Including Original Author in git commits.
I've been trying to do that recently, though as it forces me to go to the command line rather than using Atlassian Sourcetree, I've been getting other things wrong. To those people who have been dealing with commits I've managed to mess up: apologies. 1. Once someone is down as an author you don't need to add their email address; the first time you will need to get their email address 2. Akira, Aaron and I also use the -S option to GPG sign the commits. We should all be doing that, as it is the way to show who really committed the patch. Add --show-signature to the end of any git log to command to see those. 3. note that if you cherry-pick a patch into a different branch, you have to use -S in the git cherry-pick command to resign it. we should all have our GPG keys in the KEYS file, and co-sign the others in there, so that we have that mutual trust. -Steve ps: one flaw in the GPG process: if you ever revoke the key then all existing commits are considered untrusted http://steveloughran.blogspot.com/2017/10/roca-breaks-my-commit-process.html On Thu, Feb 14, 2019 at 9:12 AM Akira Ajisaka wrote: > Hi Vinay, > > I'm already doing this if I can get the original author name and the > email address in some way. > If the patch is created by git format-patch command, smart-apply-patch > --committer option can do this automatically. > Never knew that
Re: Including Original Author in git commits.
So.. if we started doing that already.. we can encourage contributors to attach formatted patch.. or create PRs. And update wiki to follow exact steps to contribute and commit. -Vinay On Thu, 14 Feb 2019, 4:54 pm Steve Loughran I've been trying to do that recently, though as it forces me to go to the > command line rather than using Atlassian Sourcetree, I've been getting > other things wrong. To those people who have been dealing with commits I've > managed to mess up: apologies. > > 1. Once someone is down as an author you don't need to add their email > address; the first time you will need to get their email address > 2. Akira, Aaron and I also use the -S option to GPG sign the commits. We > should all be doing that, as it is the way to show who really committed the > patch. Add --show-signature to the end of any git log to command to see > those. > 3. note that if you cherry-pick a patch into a different branch, you have > to use -S in the git cherry-pick command to resign it. > > we should all have our GPG keys in the KEYS file, and co-sign the others in > there, so that we have that mutual trust. > > -Steve > > ps: one flaw in the GPG process: if you ever revoke the key then all > existing commits are considered untrusted > > http://steveloughran.blogspot.com/2017/10/roca-breaks-my-commit-process.html > > > > > On Thu, Feb 14, 2019 at 9:12 AM Akira Ajisaka wrote: > > > Hi Vinay, > > > > I'm already doing this if I can get the original author name and the > > email address in some way. > > If the patch is created by git format-patch command, smart-apply-patch > > --committer option can do this automatically. > > > > Never knew that >
Re: Including Original Author in git commits.
Yetus is adding support for github PRs, we can adopt it. There's already a github a/c for it (hadoop-yetus) and a goal of adopting the new Yetus. Be great to have people involved here On Thu, Feb 14, 2019 at 1:24 PM Vinayakumar B wrote: > So.. if we started doing that already.. we can encourage contributors to > attach formatted patch.. or create PRs. > > And update wiki to follow exact steps to contribute and commit. > > -Vinay > > > On Thu, 14 Feb 2019, 4:54 pm Steve Loughran wrote: > > > I've been trying to do that recently, though as it forces me to go to the > > command line rather than using Atlassian Sourcetree, I've been getting > > other things wrong. To those people who have been dealing with commits > I've > > managed to mess up: apologies. > > > > 1. Once someone is down as an author you don't need to add their email > > address; the first time you will need to get their email address > > 2. Akira, Aaron and I also use the -S option to GPG sign the commits. We > > should all be doing that, as it is the way to show who really committed > the > > patch. Add --show-signature to the end of any git log to command to see > > those. > > 3. note that if you cherry-pick a patch into a different branch, you have > > to use -S in the git cherry-pick command to resign it. > > > > we should all have our GPG keys in the KEYS file, and co-sign the others > in > > there, so that we have that mutual trust. > > > > -Steve > > > > ps: one flaw in the GPG process: if you ever revoke the key then all > > existing commits are considered untrusted > > > > > http://steveloughran.blogspot.com/2017/10/roca-breaks-my-commit-process.html > > > > > > > > > > On Thu, Feb 14, 2019 at 9:12 AM Akira Ajisaka > wrote: > > > > > Hi Vinay, > > > > > > I'm already doing this if I can get the original author name and the > > > email address in some way. > > > If the patch is created by git format-patch command, smart-apply-patch > > > --committer option can do this automatically. > > > > > > > Never knew that > > >
Re: Including Original Author in git commits.
+1. I think formatted patches and PRs will be an improvement. I've used the git --committer thing a couple of times here without issue. Another unrelated improvement with github is handling of large changes. I really think large patches should be split up into logical subcommitts and PR's support this naturally. (We could also use something like quilt and change the JIRA process to understand patch sets but I may be the only one excited about that idea.) On Thu, Feb 14, 2019 at 5:24 AM Vinayakumar B wrote: > So.. if we started doing that already.. we can encourage contributors to > attach formatted patch.. or create PRs. > > And update wiki to follow exact steps to contribute and commit. > > -Vinay > > > On Thu, 14 Feb 2019, 4:54 pm Steve Loughran wrote: > > > I've been trying to do that recently, though as it forces me to go to the > > command line rather than using Atlassian Sourcetree, I've been getting > > other things wrong. To those people who have been dealing with commits > I've > > managed to mess up: apologies. > > > > 1. Once someone is down as an author you don't need to add their email > > address; the first time you will need to get their email address > > 2. Akira, Aaron and I also use the -S option to GPG sign the commits. We > > should all be doing that, as it is the way to show who really committed > the > > patch. Add --show-signature to the end of any git log to command to see > > those. > > 3. note that if you cherry-pick a patch into a different branch, you have > > to use -S in the git cherry-pick command to resign it. > > > > we should all have our GPG keys in the KEYS file, and co-sign the others > in > > there, so that we have that mutual trust. > > > > -Steve > > > > ps: one flaw in the GPG process: if you ever revoke the key then all > > existing commits are considered untrusted > > > > > http://steveloughran.blogspot.com/2017/10/roca-breaks-my-commit-process.html > > > > > > > > > > On Thu, Feb 14, 2019 at 9:12 AM Akira Ajisaka > wrote: > > > > > Hi Vinay, > > > > > > I'm already doing this if I can get the original author name and the > > > email address in some way. > > > If the patch is created by git format-patch command, smart-apply-patch > > > --committer option can do this automatically. > > > > > > > Never knew that > > >