Web lists-archives.com

Re: [RFC PATCH 000/194] Moving global state into the repository object




On Mon, Feb 5, 2018 at 3:51 PM, Stefan Beller <sbeller@xxxxxxxxxx> wrote:

>
> Any suggestions welcome!
>

I wouldn't say sending out this many patches is a smooth experience:

* After trying it out internally, it stopped at patch ~80 and the rate
limiter kicked in.
* Okay fine, I'll rate limit myself to be able to send out as many patches!
* Apparently I did not understand the docs correctly, and after reading the code
  a patch was made:
  https://public-inbox.org/git/20180206000743.217503-1-sbeller@xxxxxxxxxx/
  I think the sensible thing would be to error out instead of ignoring the
  relogin parameter.
  This is why patches 21/22 ff. have a weird in-reply-to setup, as I hit CTRL-C
  to stop the unlimited sending and picked up from there using the rate limited
  --relogin-delay=25 --batch-size=3
* You'll notice that at patches 99/100 ff the same weird in-reply-to appeared.
  One patch has had a missing '>' in the signoff, such that the server refused
  to cc appropriately ("Stefan" not a valid recipient)
  It would be nice if we could check for that before even sending out emails,
  but I could live with this blamed on the user. (It was my mistake)
* Patch 118 just bounced. "Message rejected."
  It is found at
https://github.com/stefanbeller/git/commit/141ba1f82c223636728a476f9acc1229f353a381

Stefan