On 10/30/23 4:02 PM, Jonathan Wakely wrote:
>
>
> On Mon, 30 Oct 2023, 19:30 Tom Honermann, <tom_at_[hidden]> wrote:
>
> Following last week's SG16 meeting that discussed, among other
> things, plans to remove the codecvt<char16, char, mbstate_t> and
> other specializations, I did some testing and was surprised to
> discover that the libstdc++ implementation of that specialization
> did not behave as I expected using gcc 11.4. I investigated and
> found I was hitting the issue at
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86419 (originally
> reported in 2018) that was fixed in January for gcc 13
> <https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=02dab998665dda0f6df31740e8897c42de3d740f>.
>
> This might be an indication that these specializations do see some
> use in the wild.
>
> Dimitrij and Jonathan, are there any details you can share
> regarding what lead to the discovery and resolution of this bug?
> The bug report indicates the issue was exposed by new tests
> written by Dimitrij. I'm curious if that testing was motivated by
> use in the community or just part of a test improvement effort.
> Jonathan, anything you can share about what motivated fixing the
> bug this year?
>
>
> It's a bug, bugs should be fixed :-)
>
> Dimitrij contributed a patch, so I saw no reason not to fix it.

Perfect. I just wanted to know if there had been some pressure to fix it
based on usage and it sounds like that is not the case.

Tom.

>
>
> Tom.
>