Date: Fri, 20 Dec 2024 16:58:35 +0200
On Fri, 20 Dec 2024 at 16:54, Jan Schultke <janschultke_at_[hidden]> wrote:
>
> > Nobody said the word "impossible", but other than that, I vehemently disagree with your suggestion that that is an "in other words, no". It's "in other words, yes".
>
> I've very specifically asked "do we NEED to, strictly speaking, from a
> technical viewpoint?". If you're not going to say "yes" but dodge the
> question and tell me how there is a potential for cyclomatic
> complexity to be increased, I can only take that as a no.
By all means. I'm done with these distortions of what I say. It's not
a "personal preference", and the reasons for it are technical.
Either pay heed to that, or don't. We don't need your proposed
facility, especially if WG14 decides to go with the approach
I and others have suggested here.
> > > We're like 10 messages into this and still only talking
> > > about "perceived complexity" and not a single problem has been brought
>
> > It's not perceived, it's measurable.
>
> Measurable how? How are you going to measure the merits of "label:
> for" vs. "for label" ? The cyclomatic complexity is identical, so that
> can't be it.
By measuring the complexity when the label is actually used as a
target of a goto. I'm well aware that the label in and of itself
has no effect on cyclomatic complexity, and said so rather clearly, by
talking about the label opening up that possibility, and that
possibility being closed off by not having a loop name be a target of a goto.
> I am aware that you cannot solve all issues with linter scripts, but
> this is one of them. You can't dodge this fact by pointing to other
> things that linter scripts aren't effective at.
Cool. I'll stop doing so, then. Linter scripts solve no problems.
Language semantics do.
> The goal here is simply to ensure that there are two different sets of
> names for goto targets and loop names. It's HIGHLY subjective whether
> that's even necessary or desirable. It's not about cyclomatic
> complexity. It's just a matter of opinion and style, and linter
> scripts can enforce opinion and style just fine.
Right, because you say so, and keep insisting so despite attempts to
explain how it's no such thing.
This discussion is over.
>
> > Nobody said the word "impossible", but other than that, I vehemently disagree with your suggestion that that is an "in other words, no". It's "in other words, yes".
>
> I've very specifically asked "do we NEED to, strictly speaking, from a
> technical viewpoint?". If you're not going to say "yes" but dodge the
> question and tell me how there is a potential for cyclomatic
> complexity to be increased, I can only take that as a no.
By all means. I'm done with these distortions of what I say. It's not
a "personal preference", and the reasons for it are technical.
Either pay heed to that, or don't. We don't need your proposed
facility, especially if WG14 decides to go with the approach
I and others have suggested here.
> > > We're like 10 messages into this and still only talking
> > > about "perceived complexity" and not a single problem has been brought
>
> > It's not perceived, it's measurable.
>
> Measurable how? How are you going to measure the merits of "label:
> for" vs. "for label" ? The cyclomatic complexity is identical, so that
> can't be it.
By measuring the complexity when the label is actually used as a
target of a goto. I'm well aware that the label in and of itself
has no effect on cyclomatic complexity, and said so rather clearly, by
talking about the label opening up that possibility, and that
possibility being closed off by not having a loop name be a target of a goto.
> I am aware that you cannot solve all issues with linter scripts, but
> this is one of them. You can't dodge this fact by pointing to other
> things that linter scripts aren't effective at.
Cool. I'll stop doing so, then. Linter scripts solve no problems.
Language semantics do.
> The goal here is simply to ensure that there are two different sets of
> names for goto targets and loop names. It's HIGHLY subjective whether
> that's even necessary or desirable. It's not about cyclomatic
> complexity. It's just a matter of opinion and style, and linter
> scripts can enforce opinion and style just fine.
Right, because you say so, and keep insisting so despite attempts to
explain how it's no such thing.
This discussion is over.
Received on 2024-12-20 14:58:49