APPROVE COMMIT NOTE: This patch has been committed # HG changeset patch # User Aidan Kehoe <kehoea(a)parhasard.net&gt; # Date 1322925235 0 # Node ID cc7f8a0e569aefdf14d87009e97ea7168be2259e # Parent e9f58d024c3ce7ea82f7061ac3ac30cd0ae26419 Accept bignums unambiguously in the syntax for object labels, lread.c. src/ChangeLog addition: 2011-12-03 Aidan Kehoe <kehoea(a)parhasard.net&gt; * lread.c (read1): Don't wrap when reading expressions that use bignums as object labels, that can lead to ambiguity and it's not actually that hard to use parse_integer() to avoid it. tests/ChangeLog addition: 2011-12-03 Aidan Kehoe <kehoea(a)parhasard.net&gt; * automated/lisp-reader-tests.el: Check that integer object labels (using the #N=... syntax) treat bignums as such, rather than as fixnums that have wrapped. diff -r e9f58d024c3c -r cc7f8a0e569a src/ChangeLog --- a/src/ChangeLog Tue Nov 29 16:19:53 2011 +0100 +++ b/src/ChangeLog Sat Dec 03 15:13:55 2011 +0000 ＠＠ -1,3 +1,10 ＠＠ +2011-12-03 Aidan Kehoe <kehoea(a)parhasard.net&gt; + + * lread.c (read1): + Don't wrap when reading expressions that use bignums as object + labels, that can lead to ambiguity and it's not actually that hard + to use parse_integer() to avoid it. + 2011-11-26 Aidan Kehoe <kehoea(a)parhasard.net&gt; * number-mp.c (bignum_to_string): diff -r e9f58d024c3c -r cc7f8a0e569a src/lread.c --- a/src/lread.c Tue Nov 29 16:19:53 2011 +0100 +++ b/src/lread.c Sat Dec 03 15:13:55 2011 +0000 ＠＠ -2697,20 +2697,29 ＠＠ case '5': case '6': case '7': case '8': case '9': /* Reader forms that can reuse previously read objects. */ { - int n = 0; - Lisp_Object found; + Lisp_Object parsed, found; + + Lstream_rewind (XLSTREAM (Vread_buffer_stream)); /* Using read_integer() here is impossible, because it - chokes on `='. Using parse_integer() is too hard. - So we simply read it in, and ignore overflows, which - is safe. */ + chokes on `='. */ while (c >= '0' && c <= '9') { - n *= 10; - n += c - '0'; + Lstream_put_ichar (XLSTREAM (Vread_buffer_stream), c); + QUIT; c = readchar (readcharfun); } - found = assq_no_quit (make_fixnum (n), Vread_objects); + + Lstream_flush (XLSTREAM (Vread_buffer_stream)); + + parsed + = parse_integer (resizing_buffer_stream_ptr + (XLSTREAM (Vread_buffer_stream)), + Lstream_byte_count (XLSTREAM + (Vread_buffer_stream)), + 10); + + found = assoc_no_quit (parsed, Vread_objects); if (c == '=') { /* #n=object returns object, but associates it with ＠＠ -2720,13 +2729,13 ＠＠ return Fsignal (Qinvalid_read_syntax, list2 (build_msg_string ("Multiply defined object label"), - make_fixnum (n))); + parsed)); } else { Lisp_Object object; - found = Fcons (make_fixnum (n), Qnil); + found = Fcons (parsed, Qnil); /* Make FOUND a placeholder for the object that will be read. (We've just consed it, and it's not visible from Lisp, so there's no possibility of ＠＠ -2751,7 +2760,7 ＠＠ return Fsignal (Qinvalid_read_syntax, list2 (build_msg_string ("Undefined symbol label"), - make_fixnum (n))); + parsed)); } return Fsignal (Qinvalid_read_syntax, list1 (build_ascstring ("#"))); diff -r e9f58d024c3c -r cc7f8a0e569a tests/ChangeLog --- a/tests/ChangeLog Tue Nov 29 16:19:53 2011 +0100 +++ b/tests/ChangeLog Sat Dec 03 15:13:55 2011 +0000 ＠＠ -1,3 +1,9 ＠＠ +2011-12-03 Aidan Kehoe <kehoea(a)parhasard.net&gt; + + * automated/lisp-reader-tests.el: + Check that integer object labels (using the #N=... syntax) treat + bignums as such, rather than as fixnums that have wrapped. + 2011-11-09 Aidan Kehoe <kehoea(a)parhasard.net&gt; Update some tests that have started failing because of some diff -r e9f58d024c3c -r cc7f8a0e569a tests/automated/lisp-reader-tests.el --- a/tests/automated/lisp-reader-tests.el Tue Nov 29 16:19:53 2011 +0100 +++ b/tests/automated/lisp-reader-tests.el Sat Dec 03 15:13:55 2011 +0000 ＠＠ -150,3 +150,8 ＠＠ (car (get-range-table #x1001 deserialized-range-table))) "checking the lisp reader handles deserialization identity, mixed")) +(when (featurep 'bignum) + (Assert (null (list-length (read (format "#%d=(1 #1=(5) 3 4 . #%d#)" + (+ most-positive-fixnum 2) + (+ most-positive-fixnum 2))))) + "checking bignum object labels don't wrap on reading")) -- ‘Iodine deficiency was endemic in parts of the UK until, through what has been described as “an unplanned and accidental public health triumph”, iodine was added to cattle feed to improve milk production in the 1930s.’ (EN Pearce, Lancet, June 2011) _______________________________________________ XEmacs-Patches mailing list XEmacs-Patches(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-patches