January 2018 - iterate-devel - mailman3.common-lisp.net

Re: Test status
by Robert Goldman 25 Jan '18

25 Jan '18

On 24 Jan 2018, at 16:55, Ilya Perminov wrote: > The latest SBCL (1.4.3) on Linux gives 2 failues: > ITERATE.TEST::BUG/WALK.2 and ITERATE.TEST::BUG/COLLECT-AT-BEGINNING. Hm. According to Attila, you should have 5 failures (which is what I have): ``` 5 out of 268 total tests failed: ITERATE.TEST::ALWAYS.FINALLY, ITERATE.TEST::NEVER.FINALLY, ITERATE.TEST::THEREIS.FINALLY, ITERATE.TEST::BUG/WALK.2, ITERATE.TEST::BUG/COLLECT-AT-BEGINNING ``` ...and it seems like there's something wrong so that I get three additional failures on Allegro. Probably don't have time to look into this now. Best, r

3 4

Re: Test status
by Robert Goldman 24 Jan '18

24 Jan '18

I think that must be using an older version of iterate, in quicklisp, because the latest version has `iterate/tests` instead (as it should). Best, r On 24 Jan 2018, at 17:12, Patrick Stein wrote: > I was going to recommend referring to cl-test-grid, but it seems that > testgrid cannot locate the iterate-tests system. > > https://common-lisp.net/project/cl-test-grid/library/iterate.html > > -- Patrick <pat(a)nklein.com> > >> On Jan 24, 2018, at 4:38 PM, Robert Goldman <rpgoldman(a)sift.info> >> wrote: >> >> On 24 Jan 2018, at 16:17, Ilya Perminov wrote: >> >> >> I think the warings are expected. A note just above them says "Note: >> These tests generate warnings involving MACROLET within Iterate". >> I am not sure why our test results are different. Maybe my SBCL >> version is too old, I'll try the latest one. >> >> Looks like these 5 failures are what I should be expecting. At the >> top of iterate-test.lisp I find this: >> >> ;; This is here to document which tests are failing at the time of >> ;; writing. It's not used directly here, but the tests of the >> ;; iterate-compat package of hu.dwim.reiterate expect it. >> (defvar *tests-expected-to-fail* >> '(always.finally >> never.finally >> thereis.finally >> bug/walk.2 >> bug/collect-at-beginning >> )) >> That suggests that the errors in code-movement.finally, >> code-movement.finally-protected and code-movement.else are unique to >> ACL and are not expected to fail. >> >> I'll try to make time to have a look at what is going on there. >> >> Best, >> r

1 0

Re: Test status
by Robert Goldman 24 Jan '18

24 Jan '18

On 24 Jan 2018, at 15:39, Ilya Perminov wrote: > Hi Robert, > > To "revert" my patch manually just revert this change: > > --- old-iterate/iterate.lisp 2018-01-15 16:49:24.171636397 -0800 > +++ new-iterate/iterate.lisp 2018-01-15 16:49:24.171636397 -0800 > @@ -3374,7 +3374,7 @@ > (if (not (and (integerp n) (> n 0))) > (clause-error "~a should be a positive integer" n-expr) > ;; Here, n is a positive integer. > - (let* ((p-i (intern-previous-info var)) > + (let* ((p-i (intern-previous-info var :next)) > (init-val (make-initial-value default default? (var-type > var))) > (temp (if (not (duplicable? init-val)) > (make-var-and-default-binding > > i.e. remove :next from (intern-previous-info var :next). > I only tested my patch on SBCL 1.3.0 and it did not break any tests. > The BUG > tests failed, but everything else passed. > > Ilya On MacOS, with SBCL 1.3.14, I get 5 failures: ``` 5 out of 268 total tests failed: ITERATE.TEST::ALWAYS.FINALLY, ITERATE.TEST::NEVER.FINALLY, ITERATE.TEST::THEREIS.FINALLY, ITERATE.TEST::BUG/WALK.2, ITERATE.TEST::BUG/COLLECT-AT-BEGINNING.; ``` I also see this: ``` ; caught STYLE-WARNING: ; undefined function: ITERATE:MULTIPLYING ; ; compilation unit finished ; Undefined function: ; ITERATE:MULTIPLYING ; caught 1 ERROR condition ; caught 1 WARNING condition ; caught 1 STYLE-WARNING condition ``` This is from loading the test system -- iterate itself loads without problems. I'm not sure why this happens: I see this: ``` iterate.lisp 2872:(defsynonym multiplying multiply) ``` and `multiplying` seems to be exported from package.lisp. The error occurs in `bug/macrolet.2` One thing I don't understand -- the test suite seems to use `:SB-RT` on SBCL, which is *not* the same as the `RT` package. Any idea what's going on here? Definitely caused me some confusion as a not-normally-SBCL user.

2 4

Test status
by Robert Goldman 24 Jan '18

24 Jan '18

I was just testing out the version of ITERATE with Ilya's new fixes. Question: what do we expect to happen when `ASDF:TEST-SYSTEM` is run on iterate? I see 8 test failures, on ACL 10.1, as follows: ``` 8 out of 267 total tests failed: ITERATE.TEST::ALWAYS.FINALLY, ITERATE.TEST::NEVER.FINALLY, ITERATE.TEST::THEREIS.FINALLY, ITERATE.TEST::CODE-MOVEMENT.ELSE, ITERATE.TEST::CODE-MOVEMENT.FINALLY, ITERATE.TEST::CODE-MOVEMENT.FINALLY-PROTECTED, ITERATE.TEST::BUG/WALK.2, ITERATE.TEST::BUG/COLLECT-AT-BEGINNING. ``` Is there some way for us to tell whether this is a regression, or if these are all known issues? I figure the tests prefixed with `BUG/` are probably known failures, but what about the other 6? Would it be possible to put a list of canonical test results on the iterate project page? Thanks!

1 1

Bug: (for X previous Y ...) does not check existence of Y
by Ilya Perminov 16 Jan '18

16 Jan '18

Hello, Iterate silently ignores "for ... previous" clauses that reference undefined variables, e.g, (iter (repeat 2) (for x previous y initially 0) (collect x)) returns (0 0). A patch to fix the bug is attached. Can someone apply it please. Thanks, Ilya

2 1