Date: Mon, 6 May 2019 17:27:57 +0200
Am 06.05.19 um 16:23 schrieb Tom Honermann:
>> My proposal N2282 adds mbstoc16s, c16stombs, mbstoc32s, c32stombs. That
>> is, non-restartable functions.
>> Non-restartable functions have a performance / code size advantage, are
>> always thread-safe, and don't read beyond the end of 0-terminated strings.
>> If there is sufficient demand for restartable ones (to be able to easily
>> handle incomplete chunks of text without needing the user to handle a
>> buffer), one could certainly consider adding those, too.
>
> Philipp, could you summarize the state of that proposal? It looks like
> it was discussed at the Pittsburgh meeting according to N2307 [1] and
> there was a request for additional implementation experience. I don't
> see an updated paper in the London pre-meeting mailing. Was this
> discussed there? Do you plan to continue pushing this forward?
>
> Tom.
The situation is still the same: WG14 wants more implementation
experience (i.e. we'd need a second implementation; I already asked on
the musl mailing list, but they didn't seem particularly interested).
Once there is a second implementation, I'd have another straw poll in WG14.
Philipp
>> My proposal N2282 adds mbstoc16s, c16stombs, mbstoc32s, c32stombs. That
>> is, non-restartable functions.
>> Non-restartable functions have a performance / code size advantage, are
>> always thread-safe, and don't read beyond the end of 0-terminated strings.
>> If there is sufficient demand for restartable ones (to be able to easily
>> handle incomplete chunks of text without needing the user to handle a
>> buffer), one could certainly consider adding those, too.
>
> Philipp, could you summarize the state of that proposal? It looks like
> it was discussed at the Pittsburgh meeting according to N2307 [1] and
> there was a request for additional implementation experience. I don't
> see an updated paper in the London pre-meeting mailing. Was this
> discussed there? Do you plan to continue pushing this forward?
>
> Tom.
The situation is still the same: WG14 wants more implementation
experience (i.e. we'd need a second implementation; I already asked on
the musl mailing list, but they didn't seem particularly interested).
Once there is a second implementation, I'd have another straw poll in WG14.
Philipp
Received on 2019-05-06 17:28:01