On Wed Nov 8, 2023 at 11:10 AM CET, Drew DeVault wrote:
> Soliciting comments to quote for a blog post on harelang.org:> "Testimonials from Hare users">> Anything hare-users has to say on this subject?
Sure! This is how I think about Hare:
At first glance, Hare looks like C with a slightly different syntax plus
some unfamiliar additions, especially tagged unions and match
expressions. After getting used to all the small differences, everything
makes sense. I can't find anything C does better.
"Feels more C than C itself".
Of course, it's a WIP, so there are rough edges and
performance is sometimes lacking
(eg reading lines from a buffered file is slower than in Go),
and as a beginner in low-level programming it's difficult to know
if it's me in the wrong or the compiler,
but the foundation is clearly so solid and KISS
that I have absolute confidence in Hare's future.
I recently implemented a small feature for Drew's shell,
I was baffled at how easy it was to enter into a foreign codebase and
make my changes. I was even able to identify a bug in a dependency!
If it had been written in some huge language like C++ or Rust,
it would have been impossible for me (non-professional programmer).
I would call attention to Hare's bootstrapping process.
Never had I ever tried compiling a language from source. Always thought
it would be a too strenuous process, better to let someone else with
more experience do it.
But then I saw Drew's talk on Hare at FOSSDEM 2023, I believe, in which
he, at the stage, built it from source.
Never had I thought that with only 3 `git clones` and 3 `make` calls one
would be able to have a (on its way to be) full fledged systems
programming language.
And so here I am today, giving my testimony on Hare, after having built
it on different distros and different architectures.
On Thu Nov 9, 2023 at 06:18 CST, Curtis Arthaud wrote:
> Of course, it's a WIP, so there are rough edges and> performance is sometimes lacking> (eg reading lines from a buffered file is slower than in Go),
That's probably a bug[1] whose fix was never merged[2]. I think the
underlying code was refactored a while back and they inherited the bug.
You can try the test case presented in that patch, the issue it's still
observable with current HEAD hare. It seems to be even worse with the
new implementation:
$ time ./newlines l.txt
8388608
real 0m32.515s
user 0m32.492s
sys 0m0.012s
$ time ./newlines l8k.txt
8388608
real 1m48.080s
user 1m48.078s
sys 0m0.000s
Unfortunately, since the underlying code changed, I don't think you can
apply the patch anymore to test if that's the issue in your case.
[1] https://lists.sr.ht/~sircmpwn/hare-dev/patches/39292
[2] https://lists.sr.ht/~sircmpwn/hare-dev/%3CCS1ZZDZK3A3Z.2YKKEJ6H2H5GJ%40debian%3E
On Thu Nov 9, 2023 at 5:58 PM UTC, Jose Lombera wrote:
>> On Thu Nov 9, 2023 at 06:18 CST, Curtis Arthaud wrote:>> > Of course, it's a WIP, so there are rough edges and> > performance is sometimes lacking> > (eg reading lines from a buffered file is slower than in Go),>> That's probably a bug[1] whose fix was never merged[2]. I think the> underlying code was refactored a while back and they inherited the bug.> You can try the test case presented in that patch, the issue it's still> observable with current HEAD hare. It seems to be even worse with the> new implementation:
looks like a good patch - i tried applying it and, as you expected, it
doesn't apply anymore. if you'd be interested in rebasing it yourself, i
can definitely review it, but if not i'll get around to writing
something similar asp
On Thu Nov 9, 2023 at 13:27 CST, Ember Sawady wrote:
> On Thu Nov 9, 2023 at 5:58 PM UTC, Jose Lombera wrote:> >> > On Thu Nov 9, 2023 at 06:18 CST, Curtis Arthaud wrote:> >> > > Of course, it's a WIP, so there are rough edges and> > > performance is sometimes lacking> > > (eg reading lines from a buffered file is slower than in Go),> >> > That's probably a bug[1] whose fix was never merged[2]. I think the> > underlying code was refactored a while back and they inherited the bug.> > You can try the test case presented in that patch, the issue it's still> > observable with current HEAD hare. It seems to be even worse with the> > new implementation:>> looks like a good patch - i tried applying it and, as you expected, it> doesn't apply anymore. if you'd be interested in rebasing it yourself, i> can definitely review it, but if not i'll get around to writing> something similar asp
Go ahead, I haven't been tracking hare-dev for a while.