7 Dec
2017
7 Dec
'17
6:34 a.m.
Could it have been having a structure field namespace per struct type?
I feel like he and I talked about that, I know that that has always been
a pet peeve of mine, I *much* prefer sbuf.st_size than whatever.size,
the former is instantly a stat structure, the later is a "whatever".
Yeah, I get it, it wasn't scalable, but I wish that there was a way to
encourage that all structure fields be uniquely named, it's awesome for
code readability.
If the kernel and common libraries could have done that, so much better.
On Thu, Dec 07, 2017 at 12:06:54AM +0000, ron minnich wrote:
A long time ago, dmr wrote about something that IIRC
BSD had done he did
not like: unlimited length identifiers in C, maybe? His argument was that
being too general was not a good thing.
Can't quite find the quote, anyone remember it? Would have been ca. 1980.
ron
--
---
Larry McVoy lm at mcvoy.com http://www.mcvoy.com/lm