Thread overview
New DConf Online 2020 Lightning Talk and a new Task Bounty
Nov 02, 2020
Mike Parker
Nov 02, 2020
Seb
Nov 02, 2020
Mike Parker
Nov 02, 2020
Seb
Nov 03, 2020
Mike Parker
November 02, 2020
We've added a "lightning talk" to the DConf Online 2020 schedule. Alexandru Militaru, whom you may remember from his DConf 2019 presentation, will give a ~15-minute talk about sil-cling, an extension to the Symmetry Integration Language (SIL). The talk is taking the 19:20 UTC slot on November 22, and Stefan Koch's talk has been moved to 19:50.

https://dconf.org/2020/online/index.html#schedule

DConf 2019 talk:
http://dconf.org/2019/talks/militaru.html

An anonymous donor has seeded a new Task Bounty with $400. The task: execute a new release of the dub-registry project.

https://www.flipcause.com/secure/cause_pdetails/MTAwNzkx

The same donor bumped up the bounty for the task to improve DLL support on Windows. It's now up to $320.

https://www.flipcause.com/secure/cause_pdetails/ODcyMDE=




November 02, 2020
On Monday, 2 November 2020 at 14:05:07 UTC, Mike Parker wrote:
>
> An anonymous donor has seeded a new Task Bounty with $400. The task: execute a new release of the dub-registry project.

Tagging was never a problem as registry is/was building fine from master:

https://github.com/dlang/dub-registry/releases/tag/v2.4.0

There are not so many tags because the registry is in maintenance mode and critical bug fixes are often directly deployed from master.

Anyhow, the real problem is that no one except Sönke has access to the registry server for deploying a release and the DLF at least in the past wasn't interested in paying for the registry server(s) themselves, s.t. they could be setup in such a way that multiple people have access to it.

I do _not_ want the bounty and while I appreciate that people want to support the D ecosystem, I do not understand why we put up a bounty to run "git tag" which would have no impact as deployment is the issue. IMHO we should address the real issues (registry lacking features + registry deployment).

A worrying side note here is that there have been many private and public mails about the root causes (three years ago the registry was crashing because the VM only had 200M of memory and a small GC leak made the server collect slightly more memory than it should). Anyhow, even without prior knowledge an email to anyone who has ever contributed code to the registry - aka its maintainers - (e.g. Sönke, Martin, WebFreak, ZombineDev, Mathias, ...) would have yielded this information - very similar to the recent LDC bounty that was done without contacting any of the LDC maintainers.
November 02, 2020
On Monday, 2 November 2020 at 15:08:43 UTC, Seb wrote:

>
> A worrying side note here is that there have been many private and public mails about the root causes (three years ago the registry was crashing because the VM only had 200M of memory and a small GC leak made the server collect slightly more memory than it should). Anyhow, even without prior knowledge an email to anyone who has ever contributed code to the registry - aka its maintainers - (e.g. Sönke, Martin, WebFreak, ZombineDev, Mathias, ...) would have yielded this information - very similar to the recent LDC bounty that was done without contacting any of the LDC maintainers.

First of all, this is nothing like the LDC situation. That was something I absolutely should have contacted someone about, and when I was made aware of my mistake I apologized for it.

My inbox is a graveyard of unanswered emails. I can't count off the top of my head the things that have been delayed or that have never happened because of that. So in the interest of getting things done, when someone wants to put a bounty on an open issue, if I judge it to be a simple contribution (i.e., something that isn't dependent on existing work, or doesn't require special consideration/permission, etc) then I'm not going to email anyone about it. The very first step on this bounty is to summarize the commits for a changelog. I don't need to contact anyone for that. The last step is to ensure that a release tag is made. That means anyone pursuing the bounty would have to contact one of the maintainers at some point. Again, it's a simple contribution, so I treated it as I would any other open issue.

My mistake was assuming that because it's an open issue, it still hasn't been resolved. Had I checked the release tags and the dates, I would have seen that a release has since been made. But I didn't, so mea culpa.

However, our conversation could have gone like this:

You: Hey, Mike, I pushed out a release in September: https://github.com/dlang/dub-registry/releases/tag/v2.4.0

Me: Oh, okay, I'll see what else he wants to put the money toward. Thanks!

So yes, I will close the bounty and ask the donor where he wants me to direct the cash.






November 02, 2020
On Monday, 2 November 2020 at 17:54:48 UTC, Mike Parker wrote:
> My mistake was assuming that because it's an open issue, it still hasn't been resolved. Had I checked the release tags and the dates,

For the dub registry releases aren't generally tagged, but often released from master.

> I would have seen that a release has since been made. But I didn't, so mea culpa.

I tagged the release as a response to your post to show how easy it is/was.

> However, our conversation could have gone like this:

I'm sorry about the rather aggressive tone.

November 03, 2020
On Monday, 2 November 2020 at 18:23:35 UTC, Seb wrote:
>
> I'm sorry about the rather aggressive tone.

Thank you.