Date: Mon, 30 Dec 2019 11:15:09 -0300
On Monday, 30 December 2019 01:34:35 -03 Tom Honermann wrote:
> That makes sense to me. When designing the version of text_view
> described in P0244 (https://wg21.link/p0244), I found no use for codecs
> in concepts or in interfaces; I specified them in terms of encoders and
> decoders. I still delivered encoder and decoder pairs as a codec that
> satisfied both encoder and decoder concepts. This seemed to work well,
> though I haven't received any specific feedback one way or another.
Hm... indeed, knowing how to get a matching pair might be useful. You don't
want to guess that you've got it right if you're doing forwards and backwards
conversions.
For those interested, this is the review link for the API we're discussing:
https://codereview.qt-project.org/c/qt/qtbase/+/285094
> That makes sense to me. When designing the version of text_view
> described in P0244 (https://wg21.link/p0244), I found no use for codecs
> in concepts or in interfaces; I specified them in terms of encoders and
> decoders. I still delivered encoder and decoder pairs as a codec that
> satisfied both encoder and decoder concepts. This seemed to work well,
> though I haven't received any specific feedback one way or another.
Hm... indeed, knowing how to get a matching pair might be useful. You don't
want to guess that you've got it right if you're doing forwards and backwards
conversions.
For those interested, this is the review link for the API we're discussing:
https://codereview.qt-project.org/c/qt/qtbase/+/285094
-- Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org Software Architect - Intel System Software Products
Received on 2019-12-30 08:17:41