I do not have a lot of CI/CD on my side, but I have a new
test case runner, which can show multi column test results
side by side, test results that come from different sources.
I already use it to show test results from these 3 systems
side by side. It was quite some work over the last 10 days
to mirgate formerly Jekejeke Prolog to the new testing framework:
1. Dogelog Player JavaScript [column name "js"]
2. Dogelog Player Python [column name "py"]
3. Formerly Jekejeke Prolog [column name "j"]
The new testing framework is ultra simple and 100% written
in Prolog. It requires a very minimal subset of Prolog. Mabye
will add further columns in the near future, like
a SWI-Prolog column or a Trealla Prolog column. The
Logtalk Testing Framework is probably a dead horse. Its
the opposite of minimal, if you first need Logtalk to
test Prolog systems. Its one more Lognonsense. I am
also phasing out the old formerly Jekejeke Prolog testing
framework. Its already phased out concerning test results
itself, what survies so far is a coverage tool. But sooner
or later this tool will also go, and be replaced by something
simpler. Also the beautifer should be replaced.
P.S.: Here are the current results, you see columns "js",
"py" and "j", its split over two test suites:
http://pages.xlog.ch/littab/doclet/docs/10_samples/04_doge_comply/reference_result/package.html
http://pages.xlog.ch/littab/doclet/docs/10_samples/04_doge_comply/frequent_result/package.html
Mostowski Collapse schrieb am Samstag, 29. Oktober 2022 um 21:18:34 UTC+2:
There is a lot of experimenting how to increase quality
assurance of Scryer Prolog. Of course continous integration (CI)
and continuous deployment (CD) could be key, especially
if you have a setup as in Scryer Prolog with many experts to
few programmers. There are like a dozend experts swirling
around and only very few programmers. So these experts
even don't have local forks, or could do some bug fixing
by themselves, which would be much more efficient. So they
need to communicate bugs and wait for new release and
do retesting, and communicate findings again. Very slow
process and not very effective because problem solving
pairs are separated. Its not that the same engineer indentifies
and fixes a problem. Here is what they are struggling with:
GitHub runners bla bla
https://github.com/mthom/scryer-prolog/discussions/966
Mostowski Collapse schrieb am Samstag, 29. Oktober 2022 um 21:10:09 UTC+2:
Aren't partial strings base on chars just simple pattern
matching. You only need to pattern match two cases:
[c1,..,cn]
[c1,..,cn|X]
Where c1,..,cn are chars and X is a variable. Its just
as easy as this proposal here:
Feature request indirect implementation of new built-ins #68
I think detection and dedicated compilation of the
following pattern would solve the issue:
Expr /\ (1 << N)
https://github.com/trealla-prolog/trealla/issues/68
Partial strings is only detection and dedicated compilation of
two patterns, still it takes already years to make Scryer Prolog
stable. What would happen if Scryer Prolog would
implement getbit/2 this way? Would this ever work?
Maybe they use the wrong framework for their interpreter
and compiler building? A framework that has what
kind of disadvantage? I don't know. What is the inherent
problem that detection and dedicated compilation becomes
a wall of problems. A big big wall, a kind of columb-wall
that might an outside believe, its an inpenetrable wall? https://en.wikipedia.org/wiki/Coulomb_barrier
LMAO!
Mostowski Collapse schrieb am Samstag, 29. Oktober 2022 um 21:02:38 UTC+2:
Does Mini Forth have partial strings? Looks like Scryer
Prolog is currently keeping itself busy with partial strings
for weeks and months.
?- read(T).
[a|b].
thread 'main' panicked at 'internal error: entered unreachable code', src/machine/machine_state_impl.rs:871:25
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
read/1 crash #1627
https://github.com/mthom/scryer-prolog/issues/1627
Quite amazing. So this idea by Ulrich Neumerkel has not
yet lead to a stable Scryer Prolog? Whats insurmountable
difficulty for partial strings?
Mostowski Collapse schrieb am Samstag, 29. Oktober 2022 um 20:47:13 UTC+2:
I think this project is dead, when was the last commit:
andrewtholt committed Mar 9, 2015 https://github.com/andrewtholt/minforth
Envy a comp.lang.prolog that isn't dead?
A. K. schrieb:
jfmc schrieb am Donnerstag, 11. August 2022 um 10:51:57 UTC+2:
Hi everyone.
Is there any way to fix the spam problem in this group?
It makes it completely unusable and hides all the interesting discussions that happened in the past.
What is the recommended forum for (multi-system) Prolog discussions nowadays?
Cheers,
Group should be closed. And open a write-only group for top spammer burs..
Or just rename the bin for him.
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)