On Monday, 17 October 2022 at 20:21:10 UTC, Dukc wrote:
On Monday, 17 October 2022 at 14:10:49 UTC, Mike Parker wrote:
Finally, I noted that I had spoken to Weka's Eyal Lotem at
DConf. He said their biggest remaining sticking point is what
happens after a move, which is what originally prompted [one
of their employees to write
DIP1014](https://github.com/dlang/DIPs/blob/master/DIPs/accepted/DIP1014.md), which introduced `opPostMove` and was accepted. It was never implemented and now never will be, thanks to [the approval of copy constructors in DIP1018](https://github.com/dlang/DIPs/blob/master/DIPs/accepted/DIP1018.md) intended to replace post blits.
Huh? First off, DIP1018 replaces postblits, not postmoves.
DIP1040, if accepted, will replace postmoves, but it's
currently post community 1, not accepted.
Yes, I know. My point is that 1018 motivated 1040: 1018 replaced
postblits with copy constructors, and 1040 replaces `opPostMove`
with move constructors for consistency. And I didn't say that
1040 has been accepted, I said 1014 was accepted. And regarding
1040, I said "we're hoping the collaboration here will result in
a stronger DIP for the next review round."
Second, the readme file at dips directory says DIP1014 is
implemented in dmd 2.088.1. Something is wrong here.
As far as I know, they aren't yet implemented. The changelog for
2.088.1 says nothing about it. So yes, that looks like it's
incorrect. I hadn't noticed that before.
Regardless, thanks for keeping us up to date.
You're welcome.
