FWIW for new subscriber, the field count is currently set at /diffusion/P/browse/master/src/applications/slowvote/controller/PhabricatorSlowvoteEditController.php;a0bf9a290a6b3f91d65d95f4101722662a17dede$178 and I have bumped it locally in my install as needed.

Sounds reasonable, thanks.

What if the command-line tools just wrote them to the db instead of the local config file? They could still be uneditable in web UI to address admin compromise concerns. Might not be worth the effort though.

I'd not seen that, no. I knew there had to a config file, but I just kinda assumed it was really only for database credentials (and that everything else was in the db).

FWIW, it looks like arc export hits this when invoked with --diff-id DXXX instead of --diff-id XXX

Could be. I guess the main thing is that it wasn't obvious. I was thinking to myself that all I had to do was clone libphutil, arcanist, and phabricator, repoint and bounce apache and then I'd be done.

You can just put the names in the URL, just field_name=value pairs. For example, here's a simplified template we use for tasks to apply changes to the production DB:

The way my install handles this is by composing task creation URL's which set a lot of the fields. Then we throw this onto a panel in the default dashboard.

Seems similar to T7514

Especially with the new requirement for a minimum svn version of 1.5, would there be any opposition to removing this check? I'm happy to make a diff for it.

Need a bugasaurus-rex token...

FWIW, it happened in the task creation workflow as well when I hit submit on this task.

FWIW, been using the following script to do these previews for an svn repo

Oh dang, that's gross.

Of course; any properly rotated token should form a loop.

You could use ^\ / SIGQUIT to force the non-graceful exit when the daemon's in the foreground and keep ^C as graceful only.

Understood. I have the daemon logs and db dumps if you think they would be helpful.

Yeah, that's the main aspect of it.

Sorry, there was a lot of background/setup story there.

Looking at the daemon logs, it looks like there were various mysql failures throughout the day (too many cnxns), but it seems like it was able to recover from those.

N.B. I couldn't decide what project to throw this into, sorry

Agreed on lack of value.

I still think that timestamping the feed events in the phab DB based on the timestamp of the commits in the repo is ideal here (although I don't know if it's feasible).

A note here that didn't get carried over from T7297, it looks like this might happen because the project gets created with an empty name and then undergoes a rename.

I even saw that task too, but the space was a distraction I guess. sorry ><

Oh, that too huh ><

Maybe this used to be a problem with older versions of svn?

Also, note that arc commit is perfectly fine with the untracked files when there are no changes on the directory itself (i.e. if I hadn't edited the svn props on .)

I am not familiar with any such restriction being imposed by subversion.

I don't want to create an audit. Just browse.

Yeah, understood, certainly wasn't suggesting that this be a priority.

Problem: items don't always show enough context to make sense of them from feeds. Example in screenshot.

The issue I'm having is exemplified in the attached screenshot.

Hmmm, based on what you merged this with, I don't think you quite understood what I was saying.

FWIW, I've been using this for what I wanted. Good for creating diffs for whole branches or between tags. Dunno if anybody else would find value in this.

Gotcha. That seems like a reasonable way to do it.

Thanks!

Don't know whether to default to triage and let somebody else assign or mark as wishlist (which I think it is).