Bug #29682 for Games-Tournament-Swiss: Upfloat (B5) ignored in c6pairs (0.09)

Sun Sep 30 13:52:24 2007 bartolin [...] gmx.de - Ticket created

Subject:	Upfloat (B5) ignored in c6pairs (0.09)
Date:	Sun, 30 Sep 2007 19:52:19 +0200
To:	bug-Games-Tournament-Swiss [...] rt.cpan.org
From:	Christian Bartolomaeus <bartolin [...] gmx.de>

Download signature.asc
application/pgp-signature 189b

Message body not shown because it is not plain text.

Hi, thanks for the new version of Games::Tournament::Swiss (0.09). I feel bad to report a problem shortly after your release. But the new version seems to ignore previous upfloats while pairing at c6pairs. With the following pairing table (it's attached as a file, also): ==== Round 3 Pairing Groups -------------------------------------------------------- Place No Opponents Roles Float Score 1 8 18,5 BW 2 2-8 1 11,10 WB U 1.5 2 12,7 BW 1.5 3 13,6 WB 1.5 4 14,9 BW 1.5 6 16,3 BW 1.5 7 17,2 WB 1.5 9 19,4 WB 1.5 9-15 5 15,8 WB 1 10 20,1 BW D 1 11 1,12 BW D 1 13 3,18 BW 1 15 5,20 BW 1 16 6,19 WB 1 17 7,14 BW 1 16 12 2,11 WB U 0.5 17-20 14 4,17 WB 0 18 8,13 WB 0 19 9,16 BW 0 20 10,15 WB 0 ==== the output of script 'pair' is: ==== Round 3: 8 (2), 1 2 3 4 6 7 9 (1.5), 5 10 11 13 15 16 17 (1), 12 (0.5), 14 18 19 20 (0), Next, Bracket 1: 8 C1, Bracket 1. Floating 8 Down, [1] 8 => [2] 8 1 2 3 4 6 7 9 Next, Bracket 2: 8 1 2 3 4 6 7 9 C1, B1,2 test: ok, no unpairables C2, x=0 C3, p=1 Heterogeneous. C4, S1 & S2: 8 & 1 2 3 4 6 7 9 C5, ordered: 8 & 1 2 3 4 6 7 9 C6, 1 paired. E1 1&8 C6others: Bracket 2's Remainder Group: 2 3 4 6 7 9 ==== 1-8 is clearly a wrong pairing, since 1 was upfloated last round. This is the third round of http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm which I use to test your module. Best regards Christian

Message body is not shown because sender requested not to inline it.

Wed Oct 10 01:11:54 2007 DRBEAN [...] cpan.org - Correspondence added

What a mess. But all of the CPAN testers are finding that 0.09 is a lot better than 0.08. :-) I guess none of the tests were testing c6pairs. I've included your pairing table and the actual pairings as a test: t/29682.t. It looks like a lot more tests like that are needed. It would have been good to find a minimal fix that would fix just this bug and nothing else, but I have gone ahead with a refactor of the indexing of the brackets. I've finished that refactor and now the test passes. I hope to upload version 0.10 in the next day or two to CPAN.

Wed Oct 10 01:11:55 2007 The RT System itself - Status changed from 'new' to 'open'

Fri Oct 12 12:15:11 2007 bartolin [...] gmx.de - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Fri, 12 Oct 2007 18:14:57 +0200
To:	Dr Bean via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Christian Bartolomaeus <bartolin [...] gmx.de>

* On 2007-10-10 Dr Bean via RT (bug-Games-Tournament-Swiss@rt.cpan.org) wrote: Show quoted text

> [...] > It would have been good to find a minimal fix that would fix just > this bug and nothing else, but I have gone ahead with a refactor of > the indexing of the brackets. > > I've finished that refactor and now the test passes. I hope to > upload version 0.10 in the next day or two to CPAN.

I just installed 0.10 and tried to pair round 3. Everything works as expected now. So I think, this bug report can be closed. Apart from this, I'm happy to report that 0.10 was able to pair the complete tournament from http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm correctly. All earlier versions failed at one or the other point. Thanks a lot.

Download signature.asc
application/pgp-signature 189b

Message body not shown because it is not plain text.

Sat Oct 13 10:52:01 2007 drbean [...] freeshell.org - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Sat, 13 Oct 2007 14:51:08 +0000
To:	Christian Bartolomaeus via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Greg Matheson <drbean [...] freeshell.org>

On Fri, 12 Oct 2007, Christian Bartolomaeus via RT wrote: Show quoted text

> 0.10 was able to pair the complete tournament from > http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm correctly. > All earlier versions failed at one or the other point.

What are the pairing tables in YAML form? They could be turned into a test to make sure it continues to pair the same way in the future.

Sat Oct 13 11:23:25 2007 bartolin [...] gmx.de - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Sat, 13 Oct 2007 17:23:02 +0200
To:	Greg Matheson via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Christian Bartolomaeus <bartolin [...] gmx.de>

Download signature.asc
application/pgp-signature 189b

Message body not shown because it is not plain text.

* On 2007-10-13 Greg Matheson via RT (bug-Games-Tournament-Swiss@rt.cpan.org) wrote: Show quoted text

> [...] > On Fri, 12 Oct 2007, Christian Bartolomaeus via RT wrote: >

> > 0.10 was able to pair the complete tournament from > > http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm correctly. > > All earlier versions failed at one or the other point.

> > What are the pairing tables in YAML form? They could be turned > into a test to make sure it continues to pair the same way in the > future.

You find the pairing tables attached for rounds 2 to 5. I guess, only the tables for the last three rounds are of interest for a test.

Message body is not shown because sender requested not to inline it.

Mon Jul 14 06:11:04 2008 bartolin [...] gmx.de - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Mon, 14 Jul 2008 12:11:23 +0200
To:	Greg Matheson via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Christian Bartolomaeus <bartolin [...] gmx.de>

Download signature.asc
application/pgp-signature 189b

Message body not shown because it is not plain text.

Hi, * On 2007-10-13 Christian Bartolomaeus (bartolin@gmx.de) wrote: Show quoted text

> [...]

> > On Fri, 12 Oct 2007, Christian Bartolomaeus via RT wrote: > >

> > > 0.10 was able to pair the complete tournament from > > > http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm correctly. > > > All earlier versions failed at one or the other point.

> > > > What are the pairing tables in YAML form? They could be turned > > into a test to make sure it continues to pair the same way in the > > future.

I went ahead and wrote a test for the complete tournament. As you recommended in the README file of 0.15 (http://search.cpan.org/src/DRBEAN/Games-Tournament-Swiss-0.15/README) I used the format of t/three.t. Actually I attach two different tests: 29682_rounds_1-4.t and 29682_rounds_1-5.t. The former works fine, while the latter doesn't. The (at least one) reason for this seems to be, that B6 isn't ignored for last round pairing for players with more than fifty percent. But I've opened a new bug report for this problem: http://rt.cpan.org/Public/Bug/Display.html?id=37669. So I think, this bug report can be closed. Best regards Christian

Message body is not shown because sender requested not to inline it.

Sat Jul 19 10:10:15 2008 drbean [...] freeshell.org - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Sat, 19 Jul 2008 14:08:41 +0000
To:	Christian Bartolomaeus via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Greg Matheson <drbean [...] freeshell.org>

On Mon, 14 Jul 2008, Christian Bartolomaeus via RT wrote: Show quoted text

> I went ahead and wrote a test for the complete tournament. As you > recommended in the README file of 0.15 > (http://search.cpan.org/src/DRBEAN/Games-Tournament-Swiss-0.15/README) > I used the format of t/three.t.

Show quoted text

> Actually I attach two different tests: 29682_rounds_1-4.t and > 29682_rounds_1-5.t. The former works fine, while the latter doesn't. > The (at least one) reason for this seems to be, that B6 isn't ignored > for last round pairing for players with more than fifty percent. But > I've opened a new bug report for this problem: > http://rt.cpan.org/Public/Bug/Display.html?id=37669.

For the present moment, perhaps we can include just rounds 1-4 and leave round 5 as a TODO test. Show quoted text

> So I think, this bug report can be closed.

Let's discuss it a bit more and only close it after including th e tests in the distribution. In the test, why not White and Black roles, rather than A and B. @Games::Tournament::Swiss::Config::roles = (qw/A B/); It would be good to put prepareTournament and runRound in a separate file some time.

Sun Jul 20 16:45:03 2008 bartolin [...] gmx.de - Correspondence added

Subject:	Re: [rt.cpan.org #29682] Upfloat (B5) ignored in c6pairs (0.09)
Date:	Sun, 20 Jul 2008 22:45:33 +0200
To:	Greg Matheson via RT <bug-Games-Tournament-Swiss [...] rt.cpan.org>
From:	Christian Bartolomaeus <bartolin [...] gmx.de>

* On 2008-07-19 Greg Matheson via RT (bug-Games-Tournament-Swiss@rt.cpan.org) wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=29682 > > [...]

> > I went ahead and wrote a test for the complete tournament. As you > > recommended in the README file of 0.15 > > (http://search.cpan.org/src/DRBEAN/Games-Tournament-Swiss-0.15/README) > > I used the format of t/three.t.

>

> > Actually I attach two different tests: 29682_rounds_1-4.t and > > 29682_rounds_1-5.t. The former works fine, while the latter doesn't. > > The (at least one) reason for this seems to be, that B6 isn't ignored > > for last round pairing for players with more than fifty percent. But > > I've opened a new bug report for this problem: > > http://rt.cpan.org/Public/Bug/Display.html?id=37669.

> > For the present moment, perhaps we can include just rounds 1-4 > and leave round 5 as a TODO test.

That sounds good to me. So, if I find some other interesting tournaments where Games::Tournament::Swiss seems to have problems, I'll try to provide two different tests: One for the part of the tournaments which works fine and one "TODO test" which includes the critical round(s). That way you can decide which test you want to add to the distribution. Show quoted text

> > So I think, this bug report can be closed.

> > Let's discuss it a bit more and only close it after including the > tests in the distribution.

Okay. Show quoted text

> In the test, why not White and Black roles, rather than A and B. > > @Games::Tournament::Swiss::Config::roles = (qw/A B/);

Indeed, that would be better. I just was lazy and took your test t/three.t as a template for my test. I will make that change to this and future tests. Show quoted text

> It would be good to put prepareTournament and runRound in a > separate file some time.

So, you are speaking about a good format to provide tests, don't you? Do you suggest to have a kind of generic test program which can be run with different data files (e.g. two files which contain the data for prepareTournament and runRound, respectively)? If you decide to do something like that I'd like to try to write a script which converts data from real tournaments (like http://www.lsvmv.de/turniere/erg/eon_2007a_paar.htm -- it was paired with SwissChess [1] and there are a lot of similar html pages for other tournaments paired with this program) to this specified format for data files for tests. That way we could easily increase the number of tests for Games::Tournament::Swiss. Or did I misunderstood your suggestion? Best regards Christian [1] http://www.swiss-chess.de/home_en.htm

Download signature.asc
application/pgp-signature 189b

Message body not shown because it is not plain text.