Attendees: Lucas (minutes), Roland (chair), Beyhan, Maxwell Rizzo, Peter, Deborah
# Accumulated updates after the release
- CarpetX (and AMReX updates): Liwei found that you can make almost all the changes so that they are backwards compatible. He submitted a PR with a patch demonstrating this. We will test this in a number of architectures to make sure it all works.
- McLachlan updates: Peter had something he was testing. Then his laptop died. He has not come around to that yet, but he will come back to it.
# Atlassian retiring Bitbucket Cloud Issues
- Zach says: He did a wrangle of clankers to help him migrate things over. He completed that about 4 days ago. He's working on a private repo for the ETK.
- What happens when user names don't match between systems? Zach did a public web search for publicly available data to find the correct user names.
- Subtlety: As soon this is moved to a public repo, everyone that has a ticket, either open or closed, will get a notification for *each ticket*. That would be a catastrophe to people like Roland and Steve.
- Who exactly is assigned to the bug? There's no way that Zach can make it as if user X opened a ticket. Zach will be listed as the ticket opener. The solution to this is to add text in the start of the ticket, saying it is ported from bitbucket, who was the original reporter, link to GH user page and other info, like original ticket number, ticket creation metadata and such.
- He proposed to use a bot, or an artificial user that is the responsible for the ticket creation (so he does not get flooded by emails regarding tickets).
- Roland points out that we need to choose a cuttof date so that a final export of the tickets is performed and no new tickets are written to the bitbucket. repo.
- The conversion will have to be done over a period of 3 days, otherwise the GH API will flag us as evil and prevent it from happening.
- ETK maintainers hace access to the scripts Zach is using for the migration.
- Bayhan asks if the migration includes the attachments in the issues. Zach says yes. Attachments will be migrated. For now the tickets are a repository and the attachment files are *in* the repo.
- Zach made it public, so that interested people can help him make the transition and test the new system.
- Roland says that it would be really nice for the ticket numbers to match between GH and Bitbucket. Zach says this may be hard: What if people deleted their tickets in bitbucket and the numbering is not as sequential as one expects. Roland says he may need to create dummy tickets so the numbers line up.
- Roland found that its also very hard to search for issues using the GH search bar and the original BB issue tag.
# Unanswered question on mailing list
- "Question about how to generate initial data with arbitrary EOS": In the body of the question, the asker specifies he is talking about BNS ID. Zach has pointed out that Sam Tootle should be the person to be talked to regarding Fuka. Also Konrad. Roland asked Beyhan to reply, he says he does not what to say. There are many ways to generate BNS ID. Roland will start the reply and others can chime in.
- "Issues configuring Simfactory for PBS cluster (Hubble) - Jobs failing to submit correctly": Beyhan directed him to fix his files.
- "Question about IllinoisGRMHD C2P failure and kuibit Psi4 extraction": Roland asked zach to chime in and ask for more time in the reply as he is busy and travelling.
# Tickets
- Looked at the most recent tickets, but had no time to discuss them in detail.