On 8/24/2018 2:45 AM, Jeff King wrote:
On Thu, Aug 23, 2018 at 10:59:55PM -0400, Jeff King wrote:
So I think we have a winner. I'll polish that up into patches and send
it out later tonight.
Oof. This rabbit hole keeps going deeper and deeper. I wrote up my
coccinelle findings separately in:
https://public-inbox.org/git/20180824064228.ga3...@sigill.intra.peff.net/
which is possibly a coccinelle bug (there I talked about oidcmp, since
it can be demonstrated with the existing transformations, but the same
thing happens with my hasheq patches). I'll wait to see how that
discussion plays out, but I do otherwise have hasheq() patches ready to
go, so it's probably not worth anybody else digging in further.
I was trying this myself yesterday, and found a doc _somewhere_ that
said what we should do is this:
@@
expression e1, e2;
@@
if (
- oidcmp(e1, e2)
+ !oideq(e1, e2)
)
(Don't forget the space before the last end-paren!)
-Stolee
