Re: [gpsd-dev] 48-hour migration heads-up

[Fred Wright]

On Thu, 30 May 2019, Eric S. Raymond wrote:
> Sanjeev Gupta <address@hidden>:
> > Eric, before you do this, and in case we cannot track down why guests
> > cannot see the repository, would it make sense to trash it, and create a
> > new one afresh?
>
> That's always an optoon if flipping the mirror switch doesn't ubjam
> whatever is jammed.
>
> This doesn'rt smell like problem with the repo itself, though.
> I think it's nore likely we've stumbled over a minor bug in a
> rarely-used part of Gitlab's web interface.

It's not just the web interface, since anonymous fetches were also not 
working.  But it indeed seems to have improved now that mirroring has been 
turned off (albeit a lot earlier than "in 48 hours" :-)).

If there isn't a global write enable on Savannah, I suggest turning off 
developers' write access for everyone but yourself (Eric), to minimize the 
possibility of pushing to Savannah by mistake.  Savannah is already 
behind, so any commits pushed there would need to be rebased for GitLab.

For some reason I got a "pipeline failure" email claiming that *I* 
had triggered it, even though it correctly listed Eric's commit as the 
source.  It looks like the CI is now green, but I wonder if it's really 
necessary to spam non-author developers with such messages.

Fred Wright