Issue 14682: Backport missing errnos to 2.7 (original) (raw)

Created on 2012-04-27 14:13 by hynek, last changed 2022-04-11 14:57 by admin. This issue is now closed.

Messages (7)
msg159458 - (view)	Author: Hynek Schlawack (hynek) *	Date: 2012-04-27 14:13
In order to fix , we need to back port at least ENOTSUP to 2.7 (presuming we don’t want to check for magic numbers). The question is, whether we should back port all/most of them when we’re at it? There doesn’t seem to be much harm to it.
msg159464 - (view)	Author: Jesús Cea Avión (jcea) *	Date: 2012-04-27 15:08
Benjamin should talk about this, but I am +1 to backport the missing DEFINEs. In fact, hardwired values are actually a bug, as reported in the linked issue.
msg159472 - (view)	Author: Antoine Pitrou (pitrou) *	Date: 2012-04-27 15:43
That's like introducing a new API, and therefore it should be limited to feature releases.
msg159474 - (view)	Author: Hynek Schlawack (hynek) *	Date: 2012-04-27 15:46
So what does that mean for ? Backport only ENOTSUP, check against “45” or “won't fix”?
msg159476 - (view)	Author: Antoine Pitrou (pitrou) *	Date: 2012-04-27 15:51
> So what does that mean for ? Backport only ENOTSUP, check > against “45” or “won't fix”? Then, backporting only ENOTSUP would be a reasonable solution. Or perhaps backport it as a private attribute (_ENOTSUP).
msg159477 - (view)	Author: Benjamin Peterson (benjamin.peterson) *	Date: 2012-04-27 15:57
I think it's okay to just backport ENOTSUP.
msg160849 - (view)	Author: Hynek Schlawack (hynek) *	Date: 2012-05-16 12:39
This one has been fixed together with #14662 in e12efebc3ba6. Thank you everyone for your input!