On Fri, Aug 12, 2016, at 12:24, Adam Bartoš wrote: > There is no buffer just on those wrapping streams because the bytes I > have are not in UTF-8. Adding one would mean a fake buffer that just > decodes and writes to the text stream. AFAIK there is no guarantee > that sys.std* objects have buffer attribute and any code relying on > that is incorrect. But I inderstand that there may be such code and we > may want to be compatible.
Yes that's what I meant, I just think it needs to be considered if we're thinking about making it (or something like it) the default python sys.std*. Maybe the decision will be that maintaining compatibility with these cases isn't important. > > - The sys.stdout obtained via streams.enable does not support > > .buffer / .buffer.raw / .detach > > - All of these objects provide a fileno() interface. > > Is this wrong? If I remember, I provide it because of some check -- > maybe in input() -- to be viewed as a stdio stream. I don't know if it's *wrong* per se (same with the no buffer/raw thing etc), I'm just concerned about the possible effects on code that is written against the current implementation. _______________________________________________ Python-ideas mailing list Python-ideas@python.org https://mail.python.org/mailman/listinfo/python-ideas Code of Conduct: http://python.org/psf/codeofconduct/
