On Thu, Nov 11, 2021 at 4:34 AM Adi Roiban <[email protected]> wrote:
> Hi, > > I am writing here to get some feedback about migration Twisted Trac > Ticket to GitHub Issues. > > In short, try to spend more time doing Twisted development and less > time developing custom tools and doing sysadmin work. > > The main reason for the migration is that we don't have enough > people/time to manage the Trac instance or improve the dev tools > around Trac. > With GitHub issues it should be easier to reuse tools created by > others for GitHub. > > Also, I think that having Pull Requests an GitHub with the GitHub own > review queue/report and a separate set of report in Trac is extra > "office" work > > We are already doing Trac authentication via Github. > > Anyone over here who is against migrating the Twisted Trac tickets to > GitHub ? > > I see Trac has about 10300 tickets and GitHub PR is now at 1672 . > We can migrate with keeping the same ID for the majority of the tickets. > > With the migration the original author of a ticket/comment is lost. > This is a security feature in GitHub. > Also, it's not trivial to get all the Trac accounts mapped to GitHub > accounts. > Maybe rather than losing this (and any other metadata) it can be preserved as non-canonical metadata. For example, maybe each issue description and comment can end with an easily machine-identifiable block that encodes the information from trac. The information may not be especially interesting in the day-to-day development of Twisted but it could potentially be interesting for certain investigations - for example, I have often been interested in finding out who made some change or comment that I could not understand so I could find them and talk to them about it. Maybe this kind of thing is already done by the migration tool anyway and you were only talking about canonical GitHub metadata, though? I've seen projects migrated to, eg, GitLab where the migration tool leaves a text blurb with original author information at the bottom, for example. > We can keep the existing Trac in read-only mode for some time. > And it can (and will be) linked from Github Issues > If *all* of the trac metadata gets encoded into GitHub *somehow* then maybe this Trac deployment doesn't have to be maintained for very long. I know this may mean extra work for the migration. However, I'd ask that you strongly weigh that cost against what's being lost if this isn't done. Twisted has decades of history at this point. That represents a massive investment from many people. Apart from the practical value that is lost if metadata about those contributions is thrown away, it also seems disrespectful to the people who put in that investment and sends a certain signal about how current and new contributions are likely to be valued by future project maintainers. Jean-Paul
_______________________________________________ Twisted mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/twisted.python.org/ Message archived at https://mail.python.org/archives/list/[email protected]/message/UKLJLXPCET7M7OHC7TOY4NQ6V4JJOT7U/ Code of Conduct: https://twisted.org/conduct
