On 01/07/22 11:58, Richard W.M. Jones wrote:
On Fri, Jan 07, 2022 at 11:24:14AM +0100, Laszlo Ersek wrote:
> On 01/07/22 11:21, Laszlo Ersek wrote:
>> On 01/06/22 16:10, Richard W.M. Jones wrote:
>>>
>>> This one looks fine, ACK
>>>
>>> No need to post v2. Fix anything I noted if you want.
>>>
>>> It'd be nice to get this into RHEL 9 by the end of this week. If you
>>> can push it later today or tomorrow then I can add it. Coincidentally
>>> I was just doing a RHEL 9 release right this moment, but I won't
>>> continue with that for now.
>>
>> I'd like you to please advise me some more under patch#6, then I'll go
>> ahead and merge the thing still today.
>
> Oh I'm sorry I see your message is from 6 Jan, and tag v1.45.96 is also
> from yesterday -- so is it still helpful if I merge this today?
If possible, yes.
OK, thank you. Here's what I'll do: I'll fix up the warts before merging
*except* the temporary value thing for caml_copy_string (). You agreed
that we should make all these temporaries explicit, and I'm happy to do
it, but that's something we should separate out. So there are two ways
to approach it: either insert a patch (dealing with the existent
instances) into this series, and then update patch#6, OR to merge
patch#6 as-is, and then post a separate patch for making the temporaries
(old and new alike) explicit. Given that what we have ATM is not a bug,
and that we'd like to merge this today, I'll take a note (actually: tag
this message after I send it) to send a separate patch for making the
temporaries explicit.
Thanks!
Laszlo
At the moment I'm doing some time-consuming benchmarking of v2v
(more on that later) so it's not a problem if it doesn't get in today.
Rich.