At 2022-12-19T20:10:25+0100, Alejandro Colomar wrote:
> On 12/19/22 17:59, G. Branden Robinson wrote:
> > [dropping Deri and groff@]
> >
> > At 2022-12-19T17:39:37+0100, Alejandro Colomar wrote:
> > > Would you mind sending a patch that I can apply with git? I could
> > > manually edit the file you attached, but I'm feeling lazy for that
> > > :\
> >
> > Damn. Our lazinesses are duelling. :P
>
> :P
>
> I actually wonder how producing such a diff was simpler than just
> copying the output of git diff...
I don't find "diff -u <tab-completion>{,.new}" very difficult.
(I often do that in repos that I mostly read as opposed to contributing
to; that way my half-baked ideas don't get in the way of a rebase pull.
Branching doesn't solve that problem, and if I git-stash something in
such a repo I'm likely to forget about it, or not think to check there.)
> BTW, I still plan releasing man-pages-6.02 in a two days, and feel
> confident enough about the string changes (modulo a few tweaks that
> I'll apply) to ship them in it. If you have any comments about them,
> please voice them :)
I have weak preference for the name "string_copying(7)" over
"string_copy(7)", but leaving the summary-description the same.
Apart from that I have no comments, apart from encouraging you in your
effort to reform common practice to something less sloppy. I think you
will get pushback from people who (1) don't appreciate how horrible the
C string library is, in part because they have mazed themselves with the
notion that the engineers at the Bell Labs CSRC were all infallible
giants who deigned to walk among us mortals for a while; and (2) would
rather wait until some total replacement solution comes along, which of
course they would oppose with at least much passion.
Nevertheless, once in a while they'll make good points. Take that
opportunity to anneal the quality of your initiative.
The standard I/O library is a disaster, too. Much more esteemed people
than I have made this point, such as Korn and Vo, who presented their
case at USENIX in 1991.[1] Of course, they did the smart thing back
then and didn't FLOSS-license it, possibly under direction from AT&T
management. Thanks to that shrewd advice, sfio.h stormed to success and
ubiquity instead of being nearly forgotten.
But even a permissive license may well have not been enough. The
average programmer will happily drink from a pool of dioxin as long as
it is a familiar one.
> > was David Wheeler (no, the other one)
>
> Didn't know of the other David Wheeler.
Well, David J. Wheeler (computer scientist) passed away in 2004, so
possibly David A. Wheeler (computer scientist) gets mistaken for him
less often now.
I've been making this joke for years. It hasn't worked once. Like the
average programmer, I learn nothing from this.
> According to wikipedia, pkg-config(1)'s initial release is also from
> 2000. So, yes, probably it wasn't widely known at the time.
It was an innocent time, when packaging was still young and exciting,
and before JavaScript people getting paid a lot to work in e-commerce
decided they could do it better than everyone else.
Regards,
Branden
[1] https://www.cs.princeton.edu/courses/archive/fall97/cs595/sfio.ps
Attachment:
signature.asc
Description: PGP signature
