[COMMIT] Don't share a counter when checking for circularity, list_merge().

Ar an dara lá déag de mí Deireadh Fómhair, scríobh Robert Pluim: From eyeballing the code, the bug seems to be that my circularity checking in list_merge() wasn’t done correctly. The below fixes a related error that I can reproduce easily; could I ask you to apply it locally and tell me if it fixes your error, too? # HG changeset patch # User Aidan Kehoe <kehoea(a)parhasard.net&gt; # Date 1286903652 -3600 # Node ID be436ac36ba448c42ea9e6305fb6ad2952be6955 # Parent dcc34e28cd84cda4cdd5bfc3b58aef06c0da78c2 Don't share a counter when checking for circularity, list_merge(). src/ChangeLog addition: 2010-10-12 Aidan Kehoe <kehoea(a)parhasard.net&gt; * fns.c (list_merge): Circularity checking here needs to be done independently for each list, they can't share a loop counter. Thank you for the bug report, Robert Pluim! tests/ChangeLog addition: 2010-10-12 Aidan Kehoe <kehoea(a)parhasard.net&gt; * automated/lisp-tests.el: Make sure circularity checking with #'merge is sane. diff -r dcc34e28cd84 -r be436ac36ba4 src/ChangeLog --- a/src/ChangeLog Sun Oct 10 12:32:38 2010 +0100 +++ b/src/ChangeLog Tue Oct 12 18:14:12 2010 +0100 ＠＠ -1,3 +1,10 ＠＠ +2010-10-12 Aidan Kehoe <kehoea(a)parhasard.net&gt; + + * fns.c (list_merge): + Circularity checking here needs to be done independently for each + list, they can't share a loop counter. Thank you for the bug + report, Robert Pluim! + 2010-09-20 Aidan Kehoe <kehoea(a)parhasard.net&gt; * lisp.h (GET_DEFUN_LISP_OBJECT): Make the NEW_GC version of this diff -r dcc34e28cd84 -r be436ac36ba4 src/fns.c --- a/src/fns.c Sun Oct 10 12:32:38 2010 +0100 +++ b/src/fns.c Tue Oct 12 18:14:12 2010 +0100 ＠＠ -2157,7 +2157,7 ＠＠ Lisp_Object l1, l2; Lisp_Object tortoises[2]; struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5; - int looped = 0; + int l1_count = 0, l2_count = 0; l1 = org_l1; l2 = org_l2; ＠＠ -2203,37 +2203,56 ＠＠ tem = l1; l1 = Fcdr (l1); org_l1 = l1; + + if (l1_count++ > CIRCULAR_LIST_SUSPICION_LENGTH) + { + if (l1_count & 1) + { + if (!CONSP (tortoises[0])) + { + mapping_interaction_error (Qmerge, tortoises[0]); + } + + tortoises[0] = XCDR (tortoises[0]); + } + + if (EQ (org_l1, tortoises[0])) + { + signal_circular_list_error (org_l1); + } + } } else { tem = l2; l2 = Fcdr (l2); org_l2 = l2; - } + + if (l2_count++ > CIRCULAR_LIST_SUSPICION_LENGTH) + { + if (l2_count & 1) + { + if (!CONSP (tortoises[1])) + { + mapping_interaction_error (Qmerge, tortoises[1]); + } + + tortoises[1] = XCDR (tortoises[1]); + } + + if (EQ (org_l2, tortoises[1])) + { + signal_circular_list_error (org_l2); + } + } + } + if (NILP (tail)) value = tem; else Fsetcdr (tail, tem); + tail = tem; - - if (++looped > CIRCULAR_LIST_SUSPICION_LENGTH) - { - if (looped & 1) - { - tortoises[0] = XCDR (tortoises[0]); - tortoises[1] = XCDR (tortoises[1]); - } - - if (EQ (org_l1, tortoises[0])) - { - signal_circular_list_error (org_l1); - } - - if (EQ (org_l2, tortoises[1])) - { - signal_circular_list_error (org_l2); - } - } } } diff -r dcc34e28cd84 -r be436ac36ba4 tests/ChangeLog --- a/tests/ChangeLog Sun Oct 10 12:32:38 2010 +0100 +++ b/tests/ChangeLog Tue Oct 12 18:14:12 2010 +0100 ＠＠ -1,3 +1,8 ＠＠ +2010-10-12 Aidan Kehoe <kehoea(a)parhasard.net&gt; + + * automated/lisp-tests.el: + Make sure circularity checking with #'merge is sane. + 2010-08-15 Aidan Kehoe <kehoea(a)parhasard.net&gt; * automated/lisp-tests.el: diff -r dcc34e28cd84 -r be436ac36ba4 tests/automated/lisp-tests.el --- a/tests/automated/lisp-tests.el Sun Oct 10 12:32:38 2010 +0100 +++ b/tests/automated/lisp-tests.el Tue Oct 12 18:14:12 2010 +0100 ＠＠ -2409,4 +2409,10 ＠＠ (Assert (not (eql '1/5 (read (prin1-to-string (intern "2/10"))))) "checking symbol named \"2/10\" not eql to ratio 1/5 on read")) +(let* ((count 0) + (list (map-into (make-list 2048 nil) #'(lambda () (decf count)))) + (expected (append list '(1)))) + (Assert (equal expected (merge 'list list '(1) #'<)) + "checking merge's circularity checks are sane")) + ;;; end of lisp-tests.el -- “Apart from the nine-banded armadillo, man is the only natural host of Mycobacterium leprae, although it can be grown in the footpads of mice.” -- Kumar & Clark, Clinical Medicine, summarising improbable leprosy research _______________________________________________ XEmacs-Patches mailing list XEmacs-Patches(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-patches