<div dir="ltr"><div dir="ltr">On Sun, 3 May 2020 at 18:44, Wookey <<a href="mailto:wookey@wookware.org">wookey@wookware.org</a>> wrote:<br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 2020-05-03 18:31 +0100, Mark Shinwell wrote:<br>
> On Sun, 3 May 2020 at 18:27, Wookey <<a href="mailto:wookey@wookware.org" target="_blank">wookey@wookware.org</a>> wrote:<br>
> <br>
> I think we should get the other years<br>
> working, otherwise it's a regression over the hg repo.<br>
> <br>
> I'm unsure that some of this is a regression, though. The sorts of things I<br>
> was referring to in my previous email (e.g. changesets with discoveries from<br>
> multiple years all at once) will be in hg too. In fact it might be worse<br>
> there, as I explicitly pulled back some of the changesets from a big ARGE merge<br>
> somewhere around 2013 so they were in a better position date-wise in the<br>
> history.<br>
<br>
True. If it's already broken in hg, then it's not a regression if it's broken in git too.<br>
<br>
I meant that things should process in git at at least one point for<br>
each year, because I'm pretty sure that was working hg. Currently<br>
that's not true for 2015, 2016, 2017, 2018, 2019 (IIRC).<br>
<br>
I agree that some data was in the wrong year in hg after 2013 so<br>
trying to fix that is an improvement. But we don't get the chance to<br>
re-arrange history very often, so it's worth some effort IMHO.<br></blockquote><div><br></div><div>(+expo-tech@ again)</div><div><br></div><div>Good progress made this evening with disentangling the CVS to svn "gap" changeset. I probably only need another evening or two, then that will be done. I'm splitting out all the renames, ARGE merges, etc. into separate changesets.</div><div><br></div><div>Mark </div></div></div>