Author Michael.Felt
Recipients Michael.Felt, dstufft, eric.araujo, taleinat
Date 2018-10-29.18:32:11
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <DB24E2D1-A9D8-40CC-9C9A-D56E31EE8899@felt.demon.nl>
In-reply-to <1540564571.03.0.788709270274.issue34897@psf.upfronthosting.co.za>
Content
> On 10/26/2018 5:36 PM, Tal Einat wrote:
> Tal Einat <taleinat@gmail.com> added the comment:
> 
> I'm not sure that the resolution currently suggested, changing compiler.set_executables(), is the right way to go.
> 
> This change to distutils is a break of backwards compatibility. Though it is a minor change, it could still break existing code.
> 
> Fixing test.support seems just as good to me in terms of code design, and better in that it is only used internally for our tests.
> 
> (BTW, instead of `elif cmd is None or (not cmd):`, you can just use `elif not cmd:`.)
Thx. One of the python bits I need to embrace.

Although - during my testing I saw that the null string "" was not being
accepted as None, but was accepted as "not cmd".

My reading error was taking None to mean a value was not initialized
while it seems to be None is a 'value' that is assigned, while ´not xxx'
can be either - never assigned or a null-length string.

This is the "ambiguity", at least for myself, that I feel I tracked down.

It is "unfortunate" if this breaks backward compatibility - IF the
backward compatibility has been based on an ambiguity. Or I again, miss
a fine-point of Python coding. Won´'t be the last time I expect.

Comments requested.

> 
> ----------
> nosy: +taleinat
> 
> _______________________________________
> Python tracker <report@bugs.python.org>
> <https://bugs.python.org/issue34897>
> _______________________________________
History
Date User Action Args
2018-10-29 18:32:11Michael.Feltsetrecipients: + Michael.Felt, taleinat, eric.araujo, dstufft
2018-10-29 18:32:11Michael.Feltlinkissue34897 messages
2018-10-29 18:32:11Michael.Feltcreate