Re: encapsulating directory operations

On 2025-05-23, James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 5/22/25 22:08, Kaz Kylheku wrote:

On 2025-05-23, James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 5/22/25 19:15, Kaz Kylheku wrote:
...

POSIX is fairly decently supported on Windows by Cygwin.

>
Ignoring for the moment the different between "fairly decently" and
"fully, Does everyone who uses Windows do so, 100% of the time, through
Cygwin? I believe not - so POSIX is not in universal use.

...

No computer speaks POSIX natively; something must be installed.

>
Yes, and POSIX could not be properly referred to as "Universal" unless
it was universally installed - which it isn't. I can't believe that this
point is being debated.

The point requires no debate if we are talking about the POSIX
environment, where if someone has such a thing, we can give them a
#!/bin/sh script that will work in that environment. (And conversely,
if they don't have such an environment, the script is of no use).

The POSIX run-time for C programs is not in the same category.
We can make programs which carry such a thing as their bundled
run-time, and transparently use it. The user doesn't see any
POSIX; just a program running on their host platform.

It's the same why we don't care that malloc or printf are not universal.
If we have a way to write hosted C for a platform and deliver programs,
that delivery mechanism will somehow ensure that malloc and printf are
there for our program, without the user having to install anything from
a third party.

--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca