October 2017 - XEmacs-commits - XEmacs Mailing List Archives

commit/XEmacs: 2 new changesets 7 years, 2 months

Bitbucket

2 new commits in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/81e6fa9d46d0/ Changeset: 81e6fa9d46d0 User: kehoea Date: 2017-10-17 13:16:04+00:00 Summary: Correct the definition of the STRING_FULLSIZE() macro, alloc.c, NEW_GC src/ChangeLog addition: 2017-10-17 Aidan Kehoe <kehoea(a)parhasard.net> * alloc.c (STRING_FULLSIZE): Correct the definition of this macro under NEW_GC, get the NEW_GC build working again. Incidentally wake up the buildbot with this commit. Affected #: 2 files diff -r 222b2a338e9daa7f4b7c12efe00859c76058e946 -r 81e6fa9d46d02bedd36f321266b3f5787055e1b0 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -7,6 +7,13 ＠＠ Removed unused file-local C variables QSsubstitute, Qsubstitute_negated from these files. +2017-10-17 Aidan Kehoe <kehoea(a)parhasard.net> + + * alloc.c (STRING_FULLSIZE): + Correct the definition of this macro under NEW_GC, get the NEW_GC + build working again. + Incidentally wake up the buildbot with this commit. + 2017-10-07 Aidan Kehoe <kehoea(a)parhasard.net> * eval.c (Feval): diff -r 222b2a338e9daa7f4b7c12efe00859c76058e946 -r 81e6fa9d46d02bedd36f321266b3f5787055e1b0 src/alloc.c --- a/src/alloc.c +++ b/src/alloc.c ＠＠ -2802,7 +2802,7 ＠＠ #ifdef NEW_GC #define STRING_FULLSIZE(size) \ - ALIGN_SIZE (FLEXIBLE_ARRAY_STRUCT_SIZEOF (Lisp_String_Direct_Data, Lisp_Object, data, (size) + 1), sizeof (Lisp_Object *)); + ALIGN_SIZE (FLEXIBLE_ARRAY_STRUCT_SIZEOF (Lisp_String_Direct_Data, Lisp_Object, data, (size) + 1), sizeof (Lisp_Object *)) #else /* not NEW_GC */ /* String blocks contain this many useful bytes. */ #define STRING_CHARS_BLOCK_SIZE \ https://bitbucket.org/xemacs/xemacs/commits/95a188885942/ Changeset: 95a188885942 User: kehoea Date: 2017-10-17 13:48:30+00:00 Summary: Handle START a bit better for string arguments, #'find, #'position src/ChangeLog addition: 2017-10-17 Aidan Kehoe <kehoea(a)parhasard.net> * sequence.c (position): Handle START a bit better, use XSTRING_ASCII_BEGIN() so we don't have to step through each character for the constant-byte-length part of the string. Affected #: 2 files diff -r 81e6fa9d46d02bedd36f321266b3f5787055e1b0 -r 95a188885942f52fda8d85b7f132163a90fd8943 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -1,3 +1,10 ＠＠ +2017-10-17 Aidan Kehoe <kehoea(a)parhasard.net> + + * sequence.c (position): + Handle START a bit better, use XSTRING_ASCII_BEGIN() so we don't + have to step through each character for the constant-byte-length + part of the string. + 2017-10-13 Aidan Kehoe <kehoea(a)parhasard.net> * doc.c: diff -r 81e6fa9d46d02bedd36f321266b3f5787055e1b0 -r 95a188885942f52fda8d85b7f132163a90fd8943 src/sequence.c --- a/src/sequence.c +++ b/src/sequence.c ＠＠ -1681,10 +1681,16 ＠＠ } else if (STRINGP (sequence)) { - Ibyte *startp = XSTRING_DATA (sequence), *cursor = startp; - Bytecount byte_len = XSTRING_LENGTH (sequence), cursor_offset = 0; + Ibyte *startp = XSTRING_DATA (sequence), *cursor; + Bytecount byte_len = XSTRING_LENGTH (sequence), cursor_offset; Lisp_Object character = Qnil; + cursor_offset = ii = (Bytecount) (min (XSTRING_ASCII_BEGIN (sequence), + starting)); + cursor = startp + cursor_offset; + /* It's probably worth making this even faster for the + non-variable-width-string case. Not now, though. */ + while (cursor_offset < byte_len && ii < ending) { if (ii >= starting) Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Remove a couple of unused file-level variables from doc.c, text.c 7 years, 2 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/222b2a338e9d/ Changeset: 222b2a338e9d User: kehoea Date: 2017-10-13 20:29:51+00:00 Summary: Remove a couple of unused file-level variables from doc.c, text.c src/ChangeLog addition: 2017-10-13 Aidan Kehoe <kehoea(a)parhasard.net> * doc.c: * doc.c (vars_of_doc): * text.c: * text.c (syms_of_text): Removed unused file-local C variables QSsubstitute, Qsubstitute_negated from these files. Affected #: 3 files diff -r 2870572bc49be089166437bb3fbaa800bcdf8d14 -r 222b2a338e9daa7f4b7c12efe00859c76058e946 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -1,3 +1,12 ＠＠ +2017-10-13 Aidan Kehoe <kehoea(a)parhasard.net> + + * doc.c: + * doc.c (vars_of_doc): + * text.c: + * text.c (syms_of_text): + Removed unused file-local C variables QSsubstitute, + Qsubstitute_negated from these files. + 2017-10-07 Aidan Kehoe <kehoea(a)parhasard.net> * eval.c (Feval): diff -r 2870572bc49be089166437bb3fbaa800bcdf8d14 -r 222b2a338e9daa7f4b7c12efe00859c76058e946 src/doc.c --- a/src/doc.c +++ b/src/doc.c ＠＠ -36,10 +36,7 ＠＠ #include "elhash.h" Lisp_Object Vinternal_doc_file_name; - -Lisp_Object QSsubstitute, Qdefvar; - -Lisp_Object Qfunction_documentation; +Lisp_Object Qdefvar, Qfunction_documentation; /* Work out what source file a function or variable came from, taking the information from the documentation file. */ ＠＠ -1399,7 +1396,4 ＠＠ Name of file containing documentation strings of built-in symbols. */ ); Vinternal_doc_file_name = Qnil; - - QSsubstitute = build_ascstring (" *substitute*"); - staticpro (&QSsubstitute); } diff -r 2870572bc49be089166437bb3fbaa800bcdf8d14 -r 222b2a338e9daa7f4b7c12efe00859c76058e946 src/text.c --- a/src/text.c +++ b/src/text.c ＠＠ -1370,7 +1370,7 ＠＠ Lisp_Object QSin_char_byte_conversion, QSin_byte_char_conversion; Lisp_Object QSin_internal_external_conversion, QSin_external_internal_conversion; -Lisp_Object /* Qfail, Qsubstitute, */ Qsubstitute_negated, Quse_private; +Lisp_Object /* Qfail, Qsubstitute, */ Quse_private; Fixnum Vchar_code_limit; ＠＠ -6116,7 +6116,6 ＠＠ DEFSUBR (Fchar_width); /* Qfail, Qsubstitute, Qsucceed in general.c */ - DEFSYMBOL (Qsubstitute_negated); /* #### what's this used for? */ DEFSYMBOL (Quse_private); #ifdef ENABLE_COMPOSITE_CHARS Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Remove #'frame-utmost-window-2 from the global namespace. 7 years, 2 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/2870572bc49b/ Changeset: 2870572bc49b User: kehoea Date: 2017-10-12 19:27:07+00:00 Summary: Remove #'frame-utmost-window-2 from the global namespace. lisp/ChangeLog addition: 2017-10-12 Aidan Kehoe <kehoea(a)parhasard.net> * frame.el (frame-utmost-window-2): Made into a label. * frame.el (frame-utmost-window-1): Use the label. Remove #'frame-utmost-window-2 from the global namespace. Affected #: 2 files diff -r d986eee8b74633db8cfee3ded33cde9de38802e3 -r 2870572bc49be089166437bb3fbaa800bcdf8d14 lisp/ChangeLog --- a/lisp/ChangeLog +++ b/lisp/ChangeLog ＠＠ -1,3 +1,9 ＠＠ +2017-10-12 Aidan Kehoe <kehoea(a)parhasard.net> + + * frame.el (frame-utmost-window-2): Made into a label. + * frame.el (frame-utmost-window-1): Use the label. + Remove #'frame-utmost-window-2 from the global namespace. + 2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> * mule/chinese.el: diff -r d986eee8b74633db8cfee3ded33cde9de38802e3 -r 2870572bc49be089166437bb3fbaa800bcdf8d14 lisp/frame.el --- a/lisp/frame.el +++ b/lisp/frame.el ＠＠ -1913,81 +1913,76 ＠＠ ;; new functions like this can be added without requiring C ;; additions. -(defun frame-utmost-window-2 (window position left-right-p major-end-p - minor-end-p) - ;; LEFT-RIGHT-P means we're looking for the leftmost or rightmost - ;; window, instead of the highest or lowest. In this case, we - ;; say that the "major axis" goes left-to-right instead of top-to- - ;; bottom. The "minor axis" always goes perpendicularly. - ;; - ;; If MAJOR-END-P is t, we're looking for a windows that abut the - ;; end (i.e. right or bottom) of the major axis, instead of the - ;; start. - ;; - ;; If MINOR-END-P is t, then we want to start counting from the - ;; end of the minor axis instead of the beginning. - ;; - ;; Here's the general idea: Imagine we're trying to count the number - ;; of windows that abut the top; call this function foo(). So, we - ;; start with the root window. If this is a vertical combination - ;; window, then foo() applied to the root window is the same as - ;; foo() applied to the first child. If the root is a horizontal - ;; combination window, then foo() applied to the root is the - ;; same as the sum of foo() applied to each of the children. - ;; Otherwise, the root window is a leaf window, and foo() is 1. - ;; Now it's clear that, each time foo() encounters a leaf window, - ;; it's encountering a different window that abuts the top. - ;; With a little examining, you can see that foo encounters the - ;; top-abutting windows in order from left to right. We can - ;; modify foo() to return the nth top-abutting window by simply - ;; keeping a global variable that is decremented each time - ;; foo() encounters a leaf window and would return 1. If the - ;; global counter gets to zero, we've encountered the window - ;; we were looking for, so we exit right away using a `throw'. - ;; Otherwise, we make sure that all normal paths return nil. - - (let (child) - (cond ((setq child (if left-right-p - (window-first-hchild window) - (window-first-vchild window))) - (if major-end-p - (while (window-next-child child) - (setq child (window-next-child child)))) - (frame-utmost-window-2 child position left-right-p major-end-p - minor-end-p)) - ((setq child (if left-right-p - (window-first-vchild window) - (window-first-hchild window))) - (if minor-end-p - (while (window-next-child child) - (setq child (window-next-child child)))) - (while child - (frame-utmost-window-2 child position left-right-p major-end-p - minor-end-p) - (setq child (if minor-end-p - (window-previous-child child) - (window-next-child child)))) - nil) - (t - (setcar position (1- (car position))) - (if (= (car position) 0) - (throw 'fhw-exit window) - nil))))) - -(defun frame-utmost-window-1 (frame position left-right-p major-end-p) - (let (minor-end-p) - (or frame (setq frame (selected-frame))) - (or position (setq position 0)) - (if (>= position 0) - (setq position (1+ position)) - (setq minor-end-p t) - (setq position (- position))) - (catch 'fhw-exit +(defun* frame-utmost-window-1 (frame position left-right-p major-end-p) + (labels + ((frame-utmost-window-2 (window position left-right-p major-end-p + minor-end-p) + ;; LEFT-RIGHT-P means we're looking for the leftmost or rightmost + ;; window, instead of the highest or lowest. In this case, we say + ;; that the "major axis" goes left-to-right instead of top-to- + ;; bottom. The "minor axis" always goes perpendicularly. + ;; + ;; If MAJOR-END-P is t, we're looking for a windows that abut the end + ;; (i.e. right or bottom) of the major axis, instead of the start. + ;; + ;; If MINOR-END-P is t, then we want to start counting from the end + ;; of the minor axis instead of the beginning. + ;; + ;; Here's the general idea: Imagine we're trying to count the number + ;; of windows that abut the top; call this function foo(). So, we + ;; start with the root window. If this is a vertical combination + ;; window, then foo() applied to the root window is the same as foo() + ;; applied to the first child. If the root is a horizontal + ;; combination window, then foo() applied to the root is the same as + ;; the sum of foo() applied to each of the children. Otherwise, the + ;; root window is a leaf window, and foo() is 1. Now it's clear + ;; that, each time foo() encounters a leaf window, it's encountering + ;; a different window that abuts the top. With a little examining, + ;; you can see that foo encounters the top-abutting windows in order + ;; from left to right. We can modify foo() to return the nth + ;; top-abutting window by simply keeping a global variable that is + ;; decremented each time foo() encounters a leaf window and would + ;; return 1. If the global counter gets to zero, we've encountered + ;; the window we were looking for, so we exit right away using a + ;; `return-from'. Otherwise, we make sure that all normal paths + ;; return nil. + (let (child) + (cond ((setq child (if left-right-p + (window-first-hchild window) + (window-first-vchild window))) + (if major-end-p + (while (window-next-child child) + (setq child (window-next-child child)))) + (frame-utmost-window-2 child position left-right-p + major-end-p minor-end-p)) + ((setq child (if left-right-p + (window-first-vchild window) + (window-first-hchild window))) + (if minor-end-p + (while (window-next-child child) + (setq child (window-next-child child)))) + (while child + (frame-utmost-window-2 child position left-right-p + major-end-p minor-end-p) + (setq child (if minor-end-p + (window-previous-child child) + (window-next-child child)))) + nil) + (t + (setcar position (1- (car position))) + (if (eql (car position) 0) + (return-from frame-utmost-window-1 window))))))) + (let (minor-end-p) + (or frame (setq frame (selected-frame))) + (or position (setq position 0)) + (if (>= position 0) + (setq position (1+ position)) + (setq minor-end-p t) + (setq position (- position))) ;; we use a cons here as a simple form of call-by-reference. ;; scheme has "boxes" for the same purpose. (frame-utmost-window-2 (frame-root-window frame) (list position) - left-right-p major-end-p minor-end-p)))) - + left-right-p major-end-p minor-end-p)))) (defun frame-highest-window (&optional frame position) "Return the highest window on FRAME which is at POSITION. ＠＠ -2036,7 +2031,6 ＠＠ If omitted, POSITION defaults to 0, i.e. the highest rightmost window. If there is no window at the given POSITION, return nil." (frame-utmost-window-1 frame position t t)) - ;; frame properties. Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Correct the sense of an assertion, old_mule_ichar_to_unicode() 7 years, 2 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/d986eee8b746/ Changeset: d986eee8b746 User: kehoea Date: 2017-10-09 07:30:37+00:00 Summary: Correct the sense of an assertion, old_mule_ichar_to_unicode() 2017-10-09 Aidan Kehoe <kehoea(a)parhasard.net> * text.c (old_mule_ichar_to_unicode): Correct the sense of an assertion here, thank you Mike Sperber! Affected #: 2 files diff -r 4ae34c21a927f5fc72839592ed99e960d7538d10 -r d986eee8b74633db8cfee3ded33cde9de38802e3 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -4,6 +4,11 ＠＠ Qunbound isn't a symbol, don't check for it once we've verified ORIGINAL_FUN is a symbol. +2017-10-09 Aidan Kehoe <kehoea(a)parhasard.net> + + * text.c (old_mule_ichar_to_unicode): + Correct the sense of an assertion here, thank you Mike Sperber! + 2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> * lisp.h: diff -r 4ae34c21a927f5fc72839592ed99e960d7538d10 -r d986eee8b74633db8cfee3ded33cde9de38802e3 src/text.c --- a/src/text.c +++ b/src/text.c ＠＠ -1703,7 +1703,8 ＠＠ old_mule_ichar_to_unicode (Ichar chr, enum converr fail) { ASSERT_VALID_ICHAR (chr); - text_checking_assert (chr < 256); + text_checking_assert (chr >= 256); /* The inline function should handle + ASCII, control-1, Latin 1. */ { int c1, c2; Lisp_Object charset; Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Don't check whether original_fun is Qunbound, we know it's a symbol, Feval() 7 years, 2 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/4ae34c21a927/ Changeset: 4ae34c21a927 User: kehoea Date: 2017-10-07 08:30:37+00:00 Summary: Don't check whether original_fun is Qunbound, we know it's a symbol, Feval() src/ChangeLog addition: 2017-10-07 Aidan Kehoe <kehoea(a)parhasard.net> * eval.c (Feval): Qunbound isn't a symbol, don't check for it once we've verified ORIGINAL_FUN is a symbol. Affected #: 2 files diff -r e3925877abce049cee4c83fcfd8a78a875ee813b -r 4ae34c21a927f5fc72839592ed99e960d7538d10 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -1,3 +1,9 ＠＠ +2017-10-07 Aidan Kehoe <kehoea(a)parhasard.net> + + * eval.c (Feval): + Qunbound isn't a symbol, don't check for it once we've verified + ORIGINAL_FUN is a symbol. + 2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> * lisp.h: diff -r e3925877abce049cee4c83fcfd8a78a875ee813b -r 4ae34c21a927f5fc72839592ed99e960d7538d10 src/eval.c --- a/src/eval.c +++ b/src/eval.c ＠＠ -3899,10 +3899,9 ＠＠ retry: /* Optimise for no indirection. */ fun = original_fun; - if (SYMBOLP (fun) && !EQ (fun, Qunbound) - && (fun = XSYMBOL (fun)->function, SYMBOLP (fun))) - { - fun = indirect_function(original_fun, 1); + if (SYMBOLP (fun) && (fun = XSYMBOL (fun)->function, SYMBOLP (fun))) + { + fun = indirect_function (original_fun, 1); } if (SUBRP (fun)) Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Drop Ben's attempt to make Big5 into one charset, unicode-internal. 7 years, 2 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/e3925877abce/ Changeset: e3925877abce User: kehoea Date: 2017-10-06 20:09:22+00:00 Summary: Drop Ben's attempt to make Big5 into one charset, unicode-internal. src/ChangeLog addition: 2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> * lisp.h: * mule-charset.c: * mule-charset.c (syms_of_mule_charset): * mule-charset.c (complex_vars_of_mule_charset): * mule-charset.c (init_mule_charset): * mule-coding.c: * mule-coding.c (byte_big5_two_byte_1_p): * mule-coding.c (big5_decode): * mule-coding.c (big5_encode): * mule-coding.c (decode_big5_char): * mule-coding.c (big5_char_to_fake_codepoint): * mule-coding.c (Fencode_big5_char): * mule-coding.c (init_mule_coding): * text.c: * unicode.c (verify_load_unicode_args): Remove the attempt to implement Big5 as one charset under unicode-internal; this makes unicode-internal builds incompatible with byte-code compiled by 21.4 and even source code, e.g. leim/quail/Punct-B5.el. lisp/ChangeLog addition: 2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> * mule/chinese.el: * mule/chinese.el (for): * mule/chinese.el (chinese-big5/list): * mule/chinese.el (find-charset): * mule/mule-category.el (predefined-category-list): * mule/mule-charset.el: Remove Ben's attempt under unicode-internal to make Big5 into one charset. This is incompatible with byte-compiled code from 21.4, which uses the hackish chinese-big5-1, chinese-big5-2 charsets with ISO-2022 escapes to encode these characters. If at some point in the future where we use UTF-8 for compiled code this could be revisited. Affected #: 10 files diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b lisp/ChangeLog --- a/lisp/ChangeLog +++ b/lisp/ChangeLog ＠＠ -1,3 +1,18 ＠＠ +2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> + + * mule/chinese.el: + * mule/chinese.el (for): + * mule/chinese.el (chinese-big5/list): + * mule/chinese.el (find-charset): + * mule/mule-category.el (predefined-category-list): + * mule/mule-charset.el: + Remove Ben's attempt under unicode-internal to make Big5 into one + charset. This is incompatible with byte-compiled code from 21.4, + which uses the hackish chinese-big5-1, chinese-big5-2 charsets + with ISO-2022 escapes to encode these characters. + If at some point in the future where we use UTF-8 for compiled + code this could be revisited. + 2017-09-30 Aidan Kehoe <kehoea(a)parhasard.net> * mule/thai.el: diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b lisp/mule/chinese.el --- a/lisp/mule/chinese.el +++ b/lisp/mule/chinese.el ＠＠ -37,11 +37,8 ＠＠ ;; Setup auto-fill-chars for charsets that should invoke auto-filling. ;; SPACE and NEWLINE are already set. -(loop for l in `(chinese-gb2312 - ,＠(if (find-charset 'chinese-big5-1) - '(chinese-big5-1 chinese-big5-2) - '(chinese-big5))) - do (put-char-table l t auto-fill-chars)) +(loop for l in '(chinese-gb2312 chinese-big5-1 chinese-big5-2) + do (put-char-table l t auto-fill-chars)) ;; CNS11643 Plane3 thru Plane7 ;; These represent more and more obscure Chinese characters. ＠＠ -111,11 +108,9 ＠＠ :list '(chinese-cns11643-1 chinese-cns11643-2 chinese-cns11643-3 chinese-cns11643-4 chinese-cns11643-5 chinese-cns11643-6 chinese-cns11643-7)) -(define-charset-tag 'chinese-big5/list - :list (if (featurep 'unicode-internal) '(chinese-big5) - '(chinese-big5-1 chinese-big5-2))) -(define-charset-tag 'chinese-gb/list - :list '(chinese-gb2312 chinese-sisheng)) +(define-charset-tag 'chinese-big5/list :list '(chinese-big5-1 chinese-big5-2)) + +(define-charset-tag 'chinese-gb/list :list '(chinese-gb2312 chinese-sisheng)) (define-charset-tag 'chinese-gb-env/list :list '(chinese-gb/list chinese-big5/list chinese-cns/list chinese/language)) ＠＠ -302,13 +297,8 ＠＠ (define-coding-system-alias 'cn-big5 'big5) -;; Need to use `compiled-when' because otherwise we will get an error when -;; compiling this file under Unicode-internal because (charset-id -;; 'chinese-big5-2) gets evaluated at compile time and fails. -(compiled-when (find-charset 'chinese-big5-1) - ;; Big5 font requires special encoding. But under Unicode-internal we - ;; have one single charset `chinese-big5', with no special encoding - ;; needed. +(when (find-charset 'chinese-big5-1) + ;; Big5 font requires special encoding. (define-ccl-program ccl-encode-big5-font `(0 ;; In: R0:chinese-big5-1 or chinese-big5-2 diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b lisp/mule/mule-category.el --- a/lisp/mule/mule-category.el +++ b/lisp/mule/mule-category.el ＠＠ -238,10 +238,8 ＠＠ ;;much better category definitions. ;;(chinese-cns11643-1 ?t "Chinese Taiwan (CNS or Big5) 2-byte character set") (chinese-cns11643-2 ?t "Chinese Taiwan (CNS or Big5) 2-byte character set") - ,＠(if (find-charset 'chinese-big5-1) - '((chinese-big5-1 ?t) - (chinese-big5-2 ?t)) - '((chinese-big5 ?t))) + (chinese-big5-1 ?t) + (chinese-big5-2 ?t) (korean-ksc5601 ?h "Hangul (Korean) 2-byte character set") (jit-ucs-charset-0 ?J "Just-in-time-allocated Unicode character") ) diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b lisp/mule/mule-charset.el --- a/lisp/mule/mule-charset.el +++ b/lisp/mule/mule-charset.el ＠＠ -1254,73 +1254,64 ＠＠ tags (cns kanji traditional-chinese chinese/language) )) -(if (featurep 'unicode-internal) - ;; We can support Big5 directly. - (make-charset - 'chinese-big5 - "Big5 (Chinese traditional)" - '(dimension 2 - ;; Big5 claims to be a 94x157 charset, but with gaps in the middle. - ;; In particular, the rows are (theoretically) indexed from A1 - FE - ;; and the columns from 40 - 7E and A1 - FE. In fact, there are gaps - ;; in the rows as well (rows C7 and C8 are missing, as well as rows - ;; FA - FE), but that appears to be due to accident -- i.e. they just - ;; ran out of chars and/or wanted to make room for expansion. Note - ;; also that the gap at C7 and C8 is due to the Level-1/Level-2 - ;; division of Big5 (see below). The 94 rows are those between - ;; A1 and FE, inclusive. The 157 columns count the sum of the columns - ;; in each disjoint set. For us, we need to use the size of the range - ;; [40, FE], which is 191. - chars (94 191) - offset (161 64) - short-name "Chinese traditional (Big5)" - long-name "Chinese traditional (Big5)" - registries ["big5.eten-0"] - unicode-map ("unicode/unicode-consortium/EASTASIA/OBSOLETE/BIG5.TXT") - tags (kanji traditional-chinese chinese/language) - )) - ;; Old Mule situation; we can only handle up to 96x96 charsets. - ;; So we split it into two charsets. According to Ken Lunde's CJKV - ;; book, Big5 itself is split into "Big Five Level 1" (rows A1-C6) - ;; and "Big Five Level 2" (rows C9-F9), with the latter containing - ;; less used characters. We split the same way then coerce the - ;; result into a 94x94 block. - (make-charset - 'chinese-big5-1 - "Frequently used part (A141-C67F) of Big5 (Chinese traditional)" - '(dimension 2 - chars 94 - final ?0 - graphic 0 - short-name "Chinese traditional (Big5), L1" - long-name "Chinese traditional (Big5) (Level-1) A141-C67F" - registries ["big5.eten-0"] - ;; no unicode map, see chinese-big5-2 - tags (kanji traditional-chinese chinese/language) - )) - (make-charset - 'chinese-big5-2 - "Less frequently used part (C940-FEFE) of Big5 (Chinese traditional)" - '(dimension 2 - chars 94 - final ?1 - graphic 0 - short-name "Chinese traditional (Big5), L2" - long-name "Chinese traditional (Big5) (Level-2) C940-FEFE" - registries ["big5.eten-0"] - ;; HACK HACK HACK! The `big5' special flag tells the internal code - ;; in Fload_unicode_mapping_table() to take codepoints out of the - ;; Big5 table, convert them to a codepoint in a "fake" chinese-big5-1 - ;; or chinese-big5-2, and store appropriately. Hence, it actually - ;; ignores the name of the charset on which the property is set and - ;; always stores in the "right" place. Correspondingly, we must set - ;; the property on big5-2, not 1, so that both charsets will be - ;; created by the time we initialize the map. - unicode-map ("unicode/unicode-consortium/EASTASIA/OBSOLETE/BIG5.TXT" - nil nil nil big5) - tags (kanji traditional-chinese chinese/language) - )) - ) +;; Big5 claims to be a 94x157 charset, but with gaps in the middle. +;; In particular, the rows are (theoretically) indexed from A1 - FE +;; and the columns from 40 - 7E and A1 - FE. In fact, there are gaps +;; in the rows as well (rows C7 and C8 are missing, as well as rows +;; FA - FE), but that appears to be due to accident -- i.e. they just +;; ran out of chars and/or wanted to make room for expansion. Note +;; also that the gap at C7 and C8 is due to the Level-1/Level-2 +;; division of Big5 (see below). The 94 rows are those between +;; A1 and FE, inclusive. The 157 columns count the sum of the columns +;; in each disjoint set. For us, we need to use the size of the range +;; [40, FE], which is 191. + +;; Old Mule situation; we can only handle up to 96x96 charsets. +;; we split it into two charsets. According to Ken Lunde's CJKV +;; book, Big5 itself is split into "Big Five Level 1" (rows A1-C6) +;; and "Big Five Level 2" (rows C9-F9), with the latter containing +;; less used characters. We split the same way then coerce the +;; result into a 94x94 block. + +;; Since escape-quoted, used for byte-compiled files and auto-saves even on +;; unicode-internal, is ISO 2022-based, it's not practical to switch, as Ben +;; attempted to do, to a single Big5 charset for unicode-internal; doing this +;; would mean byte-complied code from 21.4 would error on 21.5. + +(make-charset + 'chinese-big5-1 + "Frequently used part (A141-C67F) of Big5 (Chinese traditional)" + '(dimension 2 + chars 94 + final ?0 + graphic 0 + short-name "Chinese traditional (Big5), L1" + long-name "Chinese traditional (Big5) (Level-1) A141-C67F" + registries ["big5.eten-0"] + ;; no unicode map, see chinese-big5-2 + tags (kanji traditional-chinese chinese/language))) + +(make-charset + 'chinese-big5-2 + "Less frequently used part (C940-FEFE) of Big5 (Chinese traditional)" + '(dimension 2 + chars 94 + final ?1 + graphic 0 + short-name "Chinese traditional (Big5), L2" + long-name "Chinese traditional (Big5) (Level-2) C940-FEFE" + registries ["big5.eten-0"] + ;; HACK HACK HACK! The `big5' special flag tells the internal code + ;; in Fload_unicode_mapping_table() to take codepoints out of the + ;; Big5 table, convert them to a codepoint in a "fake" chinese-big5-1 + ;; or chinese-big5-2, and store appropriately. Hence, it actually + ;; ignores the name of the charset on which the property is set and + ;; always stores in the "right" place. Correspondingly, we must set + ;; the property on big5-2, not 1, so that both charsets will be + ;; created by the time we initialize the map. + unicode-map ("unicode/unicode-consortium/EASTASIA/OBSOLETE/BIG5.TXT" + nil nil nil big5) + tags (kanji traditional-chinese chinese/language))) ;;;;;;;;;;;;;;;;;;;;; Korean ;;;;;;;;;;;;;;;;;;;; diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -1,3 +1,25 ＠＠ +2017-10-06 Aidan Kehoe <kehoea(a)parhasard.net> + + * lisp.h: + * mule-charset.c: + * mule-charset.c (syms_of_mule_charset): + * mule-charset.c (complex_vars_of_mule_charset): + * mule-charset.c (init_mule_charset): + * mule-coding.c: + * mule-coding.c (byte_big5_two_byte_1_p): + * mule-coding.c (big5_decode): + * mule-coding.c (big5_encode): + * mule-coding.c (decode_big5_char): + * mule-coding.c (big5_char_to_fake_codepoint): + * mule-coding.c (Fencode_big5_char): + * mule-coding.c (init_mule_coding): + * text.c: + * unicode.c (verify_load_unicode_args): + Remove the attempt to implement Big5 as one charset under + unicode-internal; this makes unicode-internal builds incompatible + with byte-code compiled by 21.4 and even source code, + e.g. leim/quail/Punct-B5.el. + 2017-10-04 Aidan Kehoe <kehoea(a)parhasard.net> * text.c (old_mule_ichar_to_unicode): diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/lisp.h --- a/src/lisp.h +++ b/src/lisp.h ＠＠ -5824,12 +5824,10 ＠＠ Qchinese_cns11643_1, Qchinese_cns11643_2, #ifdef UNICODE_INTERNAL - Qchinese_big5, Qjapanese_shift_jis, -#else /* not UNICODE_INTERNAL */ +#endif /* UNICODE_INTERNAL */ Qchinese_big5_1, Qchinese_big5_2, -#endif /* UNICODE_INTERNAL */ Qchinese_sisheng, Qcomposite; diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/mule-charset.c --- a/src/mule-charset.c +++ b/src/mule-charset.c ＠＠ -60,12 +60,10 ＠＠ Lisp_Object Vcharset_japanese_jisx0212; Lisp_Object Vcharset_chinese_cns11643_1; Lisp_Object Vcharset_chinese_cns11643_2; -#ifdef UNICODE_INTERNAL -Lisp_Object Vcharset_chinese_big5; -Lisp_Object Vcharset_japanese_shift_jis; -#else Lisp_Object Vcharset_chinese_big5_1; Lisp_Object Vcharset_chinese_big5_2; +#ifdef UNICODE_INTERNAL +Lisp_Object Vcharset_japanese_shift_jis; #endif /* UNICODE_INTERNAL */ Lisp_Object Vcharset_composite; ＠＠ -197,12 +195,10 ＠＠ Qjapanese_jisx0212, Qchinese_cns11643_1, Qchinese_cns11643_2, -#ifdef UNICODE_INTERNAL - Qchinese_big5, - Qjapanese_shift_jis, -#else /* not UNICODE_INTERNAL */ Qchinese_big5_1, Qchinese_big5_2, +#ifdef UNICODE_INTERNAL + Qjapanese_shift_jis, #endif /* UNICODE_INTERNAL */ Qchinese_sisheng, Qcomposite; ＠＠ -1838,12 +1834,10 ＠＠ DEFSYMBOL (Qjapanese_jisx0212); DEFSYMBOL (Qchinese_cns11643_1); DEFSYMBOL (Qchinese_cns11643_2); -#ifdef UNICODE_INTERNAL - DEFSYMBOL (Qchinese_big5); - DEFSYMBOL (Qjapanese_shift_jis); -#else /* not UNICODE_INTERNAL */ DEFSYMBOL (Qchinese_big5_1); DEFSYMBOL (Qchinese_big5_2); +#ifdef UNICODE_INTERNAL + DEFSYMBOL (Qjapanese_shift_jis); #endif /* UNICODE_INTERNAL */ DEFSYMBOL (Qcomposite); ＠＠ -2021,17 +2015,14 ＠＠ Vcharset_chinese_cns11643_1 = Qnil; staticpro (&Vcharset_chinese_cns11643_2); Vcharset_chinese_cns11643_2 = Qnil; -#ifdef UNICODE_INTERNAL - /* We can support Shift-JIS and Big5 directly.*/ - staticpro (&Vcharset_japanese_shift_jis); - Vcharset_japanese_shift_jis = Qnil; - staticpro (&Vcharset_chinese_big5); - Vcharset_chinese_big5 = Qnil; -#else staticpro (&Vcharset_chinese_big5_1); Vcharset_chinese_big5_1 = Qnil; staticpro (&Vcharset_chinese_big5_2); Vcharset_chinese_big5_2 = Qnil; +#ifdef UNICODE_INTERNAL + /* We can support Shift-JIS directly.*/ + staticpro (&Vcharset_japanese_shift_jis); + Vcharset_japanese_shift_jis = Qnil; #endif /* UNICODE_INTERNAL */ #ifdef ENABLE_COMPOSITE_CHARS ＠＠ -2144,12 +2135,10 ＠＠ Vcharset_japanese_jisx0212 = Fget_charset (Qjapanese_jisx0212); Vcharset_chinese_cns11643_1 = Fget_charset (Qchinese_cns11643_1); Vcharset_chinese_cns11643_2 = Fget_charset (Qchinese_cns11643_2); -#ifdef UNICODE_INTERNAL - /* We can support Shift-JIS and Big5 directly.*/ - Vcharset_japanese_shift_jis = Fget_charset (Qjapanese_shift_jis); - Vcharset_chinese_big5 = Fget_charset (Qchinese_big5); -#else Vcharset_chinese_big5_1 = Fget_charset (Qchinese_big5_1); Vcharset_chinese_big5_2 = Fget_charset (Qchinese_big5_2); +#ifdef UNICODE_INTERNAL + /* We can support Shift-JIS directly.*/ + Vcharset_japanese_shift_jis = Fget_charset (Qjapanese_shift_jis); #endif /* UNICODE_INTERNAL */ } diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/mule-coding.c --- a/src/mule-coding.c +++ b/src/mule-coding.c ＠＠ -898,8 +898,6 ＠＠ /* BIG5 (used for Mandarin in Taiwan). */ DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (big5); -#ifndef UNICODE_INTERNAL - /* BIG5 is a coding system encoding two character sets: ASCII and Big5. An ASCII character is encoded as is. Big5 is a two-byte character set and is encoded in two-byte. ＠＠ -991,8 +989,6 ＠＠ b2 += b2 < 0x3F ? 0x40 : 0x62; \ } while (0) -#endif /* not UNICODE_INTERNAL */ - inline static int byte_big5_two_byte_1_p (int c) { ＠＠ -1024,17 +1020,11 ＠＠ /* Previous character was first byte of Big5 char. */ if (byte_big5_two_byte_2_p (c)) { -#ifdef UNICODE_INTERNAL - non_ascii_charset_codepoint_to_dynarr - (Vcharset_chinese_big5, data->ch, c, dst, - CONVERR_USE_PRIVATE); -#else /* not UNICODE_INTERNAL */ Lisp_Object charset; int b1, b2; DECODE_BIG5 (data->ch, c, charset, b1, b2); non_ascii_charset_codepoint_to_dynarr (charset, b1, b2, dst, CONVERR_USE_PRIVATE); -#endif /* UNICODE_INTERNAL */ } else { ＠＠ -1084,13 +1074,6 ＠＠ } ichar_to_charset_codepoint (ich, Vbig5_precedence, &charset, &c1, &c2, CONVERR_FAIL); -#ifdef UNICODE_INTERNAL - if (EQ (charset, Vcharset_chinese_big5)) - { - Dynarr_add (dst, c1); - Dynarr_add (dst, c2); - } -#else /* not UNICODE_INTERNAL */ if (EQ (charset, Vcharset_chinese_big5_1) || EQ (charset, Vcharset_chinese_big5_2)) { ＠＠ -1099,7 +1082,6 ＠＠ Dynarr_add (dst, b1); Dynarr_add (dst, b2); } -#endif /* UNICODE_INTERNAL */ else { handle_standard_encoding_error (str, src, dst); ＠＠ -1131,10 +1113,6 ＠＠ static Ichar decode_big5_char (int b1, int b2) { -#ifdef UNICODE_INTERNAL - return charset_codepoint_to_ichar (Vcharset_chinese_big5, b1, b2, - CONVERR_FAIL); -#else /* not UNICODE_INTERNAL */ if (byte_big5_two_byte_1_p (b1) && byte_big5_two_byte_2_p (b2)) { ＠＠ -1146,11 +1124,8 ＠＠ } else return -1; -#endif /* UNICODE_INTERNAL */ } -#ifndef UNICODE_INTERNAL - void big5_char_to_fake_codepoint (int b1, int b2, Lisp_Object *charset, int *c1, int *c2) ＠＠ -1168,8 +1143,6 ＠＠ } } -#endif /* not UNICODE_INTERNAL */ - DEFUN ("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /* Convert Big Five character codes in CODE into a character. CODE is a cons of two integers specifying the codepoints in Big Five. ＠＠ -1208,12 +1181,6 ＠＠ CHECK_CHAR_COERCE_INT (character); ichar_to_charset_codepoint (XCHAR (character), Vbig5_precedence, &charset, &c1, &c2, CONVERR_FAIL); -#ifdef UNICODE_INTERNAL - if (EQ (charset, Vcharset_chinese_big5)) - { - return Fcons (make_fixnum (c1), make_fixnum (c2)); - } -#else /* not UNICODE_INTERNAL */ if (EQ (charset, Vcharset_chinese_big5_1) || EQ (charset, Vcharset_chinese_big5_2)) { ＠＠ -1221,7 +1188,6 ＠＠ ENCODE_BIG5 (charset, c1, c2, b1, b2); return Fcons (make_fixnum (b1), make_fixnum (b2)); } -#endif /* UNICODE_INTERNAL */ else return Qnil; } ＠＠ -4187,14 +4153,7 ＠＠ simple_convert_predence_list_to_array (list3 (Vcharset_japanese_jisx0208, Vcharset_japanese_jisx0208_1978, Vcharset_katakana_jisx0201)); - -#ifdef UNICODE_INTERNAL - Vbig5_precedence = - simple_convert_predence_list_to_array - (list1 (Vcharset_chinese_big5)); -#else /* not UNICODE_INTERNAL */ Vbig5_precedence = simple_convert_predence_list_to_array (list2 (Vcharset_chinese_big5_1, Vcharset_chinese_big5_2)); -#endif /* UNICODE_INTERNAL */ } diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/text.c --- a/src/text.c +++ b/src/text.c ＠＠ -5521,14 +5521,11 ＠＠ All codes above are two decimal numbers except for Big Five and ANSI Z39.64, which we don't support. We add 32 to each of the decimal numbers. -Note that in an old-Mule world (see below), Big Five is split in a rather -hackish fashion into two charsets, `chinese-big5-1' and `chinese-big5-2', -due to its excessive size (94x157), with the first codepoint in the range -0xA1 to 0xFE and the second in the range 0x40 to 0x7E or 0xA1 to 0xFE. -`decode-big5-char' is used to generate the char from its codes, and -`encode-big5-char' extracts the codes. This hack doesn't exist with -Unicode-internal, and hence the expression (make-char 'chinese-big5 164 116) -could be used. +Big Five is split in a rather hackish fashion into two charsets, +`chinese-big5-1' and `chinese-big5-2', due to its excessive size (94x157), +with the first codepoint in the range 0xA1 to 0xFE and the second in the range +0x40 to 0x7E or 0xA1 to 0xFE. `decode-big5-char' is used to generate the char +from its codes, and `encode-big5-char' extracts the codes. Note that there are three different internal formats for characters: diff -r 19876e056bb7630f1854719c8b037a1f2fb691e8 -r e3925877abce049cee4c83fcfd8a78a875ee813b src/unicode.c --- a/src/unicode.c +++ b/src/unicode.c ＠＠ -2248,10 +2248,8 ＠＠ { if (EQ (elt, Qignore_first_column)) *flags_out |= LOAD_UNICODE_IGNORE_FIRST_COLUMN; -#ifndef UNICODE_INTERNAL else if (EQ (elt, Qbig5)) *flags_out |= LOAD_UNICODE_BIG5; -#endif /* not UNICODE_INTERNAL */ else invalid_constant ("Unrecognized `load-unicode-mapping-table' flag", elt); ＠＠ -2324,9 +2322,7 ＠＠ int flgs; int stage; int to_unicode_min_val[256], to_unicode_max_val[256]; -#ifndef UNICODE_INTERNAL int big5_other_unicode_min_val[256], big5_other_unicode_max_val[256]; -#endif int i; /* This may be called to autoload the Unicode tables, from a function ＠＠ -2338,7 +2334,6 ＠＠ verify_load_unicode_args (filename, start, end, offset, flags, &st, &en, &of, &flgs); -#ifndef UNICODE_INTERNAL if (flgs & LOAD_UNICODE_BIG5) { /* At this point the charsets haven't been initialized ＠＠ -2353,7 +2348,6 ＠＠ big5_other_unicode_max_val[i] = 0; } } -#endif /* not UNICODE_INTERNAL */ for (i = 0; i < 256; i++) { ＠＠ -2482,7 +2476,6 ＠＠ cp1high = cp1 >> 8; cp1low = cp1 & 255; -#ifndef UNICODE_INTERNAL if (flgs & LOAD_UNICODE_BIG5) { big5_char_to_fake_codepoint (cp1high, cp1low, ＠＠ -2509,7 +2502,6 ＠＠ set_unicode_conversion (cp2, charset, c1, c2); } else -#endif /* not UNICODE_INTERNAL */ { int l1, l2, h1, h2; c1 = cp1high, c2 = cp1low; ＠＠ -2520,9 +2512,7 ＠＠ goto out_of_range; } -#ifndef UNICODE_INTERNAL do_it: -#endif if (stage == 0) { if (c2 < to_unicode_min_val[c1]) Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/XEmacs: kehoea: Be more realistic, #'char-to-unicode, #'make-char 7 years, 3 months

Bitbucket

1 new commit in XEmacs: https://bitbucket.org/xemacs/xemacs/commits/19876e056bb7/ Changeset: 19876e056bb7 User: kehoea Date: 2017-10-04 23:08:17+00:00 Summary: Be more realistic, #'char-to-unicode, #'make-char 2017-10-04 Aidan Kehoe <kehoea(a)parhasard.net> * text.c (old_mule_ichar_to_unicode): Assert we haven't been handed an Ichar less than 256 in value. * text.c (unicode_internal_handle_bad_ichar_to_unicode): New. Handle a value passed to the inline ichar_to_unicode(), if CHR is not an official Unicode character and FAIL is not CONVERR_USE_PRIVATE. * text.c (Fmake_char): Make this API compatible with GNU and with pre-unicode-internal, return a private character by default if HANDLE-ERROR is nil. * text.c (Fchar_to_unicode): Return nil if HANDLE-ERROR is FAIL or nil here. * text.c (Funicode_to_char): * text.c (Fchar_to_charset_codepoint): * text.c (Fchar_charset): * text.c (Fsplit_char): Update the docstrings in this three functions. * text.h (ichar_to_unicode): Make the implementation of this a bit more realistic with UNICODE_INTERNAL, use the new unicode_internal_handle_bad_ichar_to_unicode(). Affected #: 3 files diff -r 8df05e6e201fbbd618c7127c762bf954028ceec3 -r 19876e056bb7630f1854719c8b037a1f2fb691e8 src/ChangeLog --- a/src/ChangeLog +++ b/src/ChangeLog ＠＠ -1,3 +1,26 ＠＠ +2017-10-04 Aidan Kehoe <kehoea(a)parhasard.net> + + * text.c (old_mule_ichar_to_unicode): + Assert we haven't been handed an Ichar less than 256 in value. + * text.c (unicode_internal_handle_bad_ichar_to_unicode): New. + Handle a value passed to the inline ichar_to_unicode(), if CHR is + not an official Unicode character and FAIL is not + CONVERR_USE_PRIVATE. + * text.c (Fmake_char): + Make this API compatible with GNU and with pre-unicode-internal, + return a private character by default if HANDLE-ERROR is nil. + * text.c (Fchar_to_unicode): + Return nil if HANDLE-ERROR is FAIL or nil here. + * text.c (Funicode_to_char): + * text.c (Fchar_to_charset_codepoint): + * text.c (Fchar_charset): + * text.c (Fsplit_char): + Update the docstrings in this three functions. + * text.h (ichar_to_unicode): + Make the implementation of this a bit more realistic with + UNICODE_INTERNAL, use the new + unicode_internal_handle_bad_ichar_to_unicode(). + 2017-09-30 Aidan Kehoe <kehoea(a)parhasard.net> * text.c (old_mule_non_ascii_valid_ichar_p): diff -r 8df05e6e201fbbd618c7127c762bf954028ceec3 -r 19876e056bb7630f1854719c8b037a1f2fb691e8 src/text.c --- a/src/text.c +++ b/src/text.c ＠＠ -1703,11 +1703,7 ＠＠ old_mule_ichar_to_unicode (Ichar chr, enum converr fail) { ASSERT_VALID_ICHAR (chr); - - /* This shortcut depends on the representation of an Ichar, see text.c. */ - if (chr < 256) - return (int) chr; - + text_checking_assert (chr < 256); { int c1, c2; Lisp_Object charset; ＠＠ -1862,6 +1858,39 ＠＠ #endif /* not UNICODE_INTERNAL */ +int +unicode_internal_handle_bad_ichar_to_unicode (Ichar chr, enum converr fail) +{ + ASSERT_VALID_ICHAR (chr); + text_checking_assert (fail != CONVERR_USE_PRIVATE); + text_checking_assert (!valid_unicode_codepoint_p (chr, + UNICODE_OFFICIAL_ONLY)); + switch (fail) + { + case CONVERR_FAIL: + return -1; + case CONVERR_ERROR: + if (chr > UNICODE_OFFICIAL_MAX) + { + args_out_of_range_3 (make_fixnum (chr), Qzero, + make_fixnum (UNICODE_OFFICIAL_MAX)); + } + if (chr >= FIRST_UTF_16_SURROGATE) + { + args_out_of_range_3 (make_fixnum (chr), Qzero, + make_fixnum (FIRST_UTF_16_SURROGATE)); + } + /* FALLTHROUGH */ + case CONVERR_ABORT: + case CONVERR_USE_PRIVATE: + default: + ABORT (); + case CONVERR_SUCCEED: + case CONVERR_SUBSTITUTE: + return 0xFFFD; + } +} + #endif /* MULE */ /* Take a possibly invalid Ichar value (must be >= 0) and move upwards as ＠＠ -5422,11 +5451,11 ＠＠ to translate. Currently, this happens only with Unicode-internal (see below): -nil or `fail' Return nil -`abort' Signal an error -`succeed' Same as `substitute' -`substitute' Substitute the Unicode replacement char (0xFFFD) -`use-private' Encode using private Unicode space +`use-private' or `nil' Encode using private Unicode space +`fail' Return nil +`error' Signal an error +`succeed' Same as `substitute' +`substitute' Substitute the Unicode replacement char (0xFFFD) Each octet should be in the range corresponding to the offset and size for that dimension, as defined in the charset. For a typical one-dimensional ＠＠ -5602,10 +5631,16 ＠＠ */ (charset, octet1, octet2, handle_error)) { - enum converr fail = decode_handle_error (handle_error, 1); + enum converr fail; int a1, a2; Ichar ch; + if (NILP (handle_error)) + { + handle_error = Quse_private; + } + + fail = decode_handle_error (handle_error, 1); charset = get_external_charset_codepoint (charset, octet1, octet2, &a1, &a2, 1); ch = charset_codepoint_to_ichar (charset, a1, a2, fail); ＠＠ -5622,7 +5657,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute the Unicode replacement char (0xFFFD) `use-private' Encode using private Unicode space ＠＠ -5630,9 +5665,18 ＠＠ (character, handle_error)) { enum converr fail = decode_handle_error (handle_error, 1); + INT_32_BIT result; CHECK_CHAR_COERCE_INT (character); - return make_fixnum (ichar_to_unicode (XCHAR (character), fail)); + + result = ichar_to_unicode (XCHAR (character), fail); + + if (result < 0) + { + return Qnil; + } + + return make_fixnum (result); } DEFUN ("unicode-to-char", Funicode_to_char, 1, 3, 0, /* ＠＠ -5684,7 +5728,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute a '?' character */ ＠＠ -5727,7 +5771,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute a '?' character */ ＠＠ -5761,7 +5805,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute the Unicode replacement char (0xFFFD) `use-private' Encode using private Unicode space ＠＠ -5799,7 +5843,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute a '?' character */ ＠＠ -5849,7 +5893,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute a '?' character */ ＠＠ -5928,7 +5972,7 ＠＠ HANDLE-ERROR controls error behavior: nil or `fail' Return nil -`abort' Signal an error +`error' Signal an error `succeed' Same as `substitute' `substitute' Substitute a '?' character */ diff -r 8df05e6e201fbbd618c7127c762bf954028ceec3 -r 19876e056bb7630f1854719c8b037a1f2fb691e8 src/text.h --- a/src/text.h +++ b/src/text.h ＠＠ -522,15 +522,17 ＠＠ #define LAST_TRAILING_BYTE 0xFF #endif -#ifndef UNICODE_INTERNAL -MODULE_API int old_mule_non_ascii_valid_ichar_p (Ichar ch); +#ifdef UNICODE_INTERNAL +MODULE_API INT_32_BIT unicode_internal_handle_bad_ichar_to_unicode (Ichar, + enum + converr); +#elif defined (MULE) +MODULE_API INT_32_BIT old_mule_non_ascii_valid_ichar_p (Ichar ch); #endif -/* Return whether the given Ichar is valid. - */ - +/* Return whether the given Ichar is valid. */ DECLARE_INLINE_HEADER ( -int +Boolint valid_ichar_p (Ichar ch) ) { ＠＠ -826,12 +828,30 ＠＠ Return value will be -1 if cannot convert. */ DECLARE_INLINE_HEADER ( int -ichar_to_unicode (Ichar chr, enum converr USED_IF_OLD_MULE (fail)) +ichar_to_unicode (Ichar chr, enum converr USED_IF_MULE (fail)) ) { ASSERT_VALID_ICHAR (chr); -#if defined (MULE) && !defined (UNICODE_INTERNAL) + /* This shortcut depends on the representation of an Ichar, see text.c. */ + if (chr < 256) + { + return (int) chr; + } + +#ifdef UNICODE_INTERNAL + if (fail == CONVERR_USE_PRIVATE) + { + return (int) chr; + } + + if (valid_unicode_codepoint_p (chr, UNICODE_OFFICIAL_ONLY)) + { + return (int) chr; + } + + return unicode_internal_handle_bad_ichar_to_unicode (chr, fail); +#elif defined (MULE) return old_mule_ichar_to_unicode (chr, fail); #else /* Unicode-internal or non-Mule */ Repository URL: https://bitbucket.org/xemacs/xemacs/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/cc-mode: acm: Fontify untyped function declarations in C Mode correctly. 7 years, 3 months

Bitbucket

1 new commit in cc-mode: https://bitbucket.org/xemacs/cc-mode/commits/57b32f709a4f/ Changeset: 57b32f709a4f User: acm Date: 2017-10-04 16:58:54+00:00 Summary: Fontify untyped function declarations in C Mode correctly. Also correct two bugs where deleting WS at a BOL could leave an untyped function declaration unfontified. * cc-engine.el (c-find-decl-spots): Don't set the flag "top-level" when we're in a macro. (c-forward-decl-or-cast-1): Recognize top-level "foo(bar)" or "foo()" in C Mode as a implicitly typed function declaration. (c-just-after-func-arglist-p): Don't get confused by "defined (foo)" inside a macro. It's not a function plus arglist. * cc-langs.el (c-cpp-expr-functions-key): New defconst and defvar. * cc-mode.el (c-fl-decl-end): After c-forward-declarator, move over any following parenthesis expression (i.e. parameter list). (c-change-expand-fl-region): When c-new-END is at a BOL, include that line in the returned region, to cope with deletions at column 0. Affected #: 3 files diff -r e39b785df77e0a0f89c55aa73e042cc6cc499028 -r 57b32f709a4f3f80c78071d67fec82b4cc629878 cc-engine.el --- a/cc-engine.el +++ b/cc-engine.el ＠＠ -134,7 +134,7 ＠＠ ;; ;; 'c-not-decl ;; Put on the brace which introduces a brace list and on the commas -;; which separate the element within it. +;; which separate the elements within it. ;; ;; 'c-awk-NL-prop ;; Used in AWK mode to mark the various kinds of newlines. See ＠＠ -5413,8 +5413,8 ＠＠ (min c-bs-cache-limit pos))) (defun c-update-brace-stack (stack from to) - ;; Give a brace-stack which has the value STACK at position FROM, update it - ;; to it's value at position TO, where TO is after (or equal to) FROM. + ;; Given a brace-stack which has the value STACK at position FROM, update it + ;; to its value at position TO, where TO is after (or equal to) FROM. ;; Return a cons of either TO (if it is outside a literal) and this new ;; value, or of the next position after TO outside a literal and the new ;; value. ＠＠ -5659,11 +5659,13 ＠＠ ;; Call CFD-FUN for each possible spot for a declaration, cast or ;; label from the point to CFD-LIMIT. ;; - ;; CFD-FUN is called with point at the start of the spot. It's passed two + ;; CFD-FUN is called with point at the start of the spot. It's passed three ;; arguments: The first is the end position of the token preceding the spot, ;; or 0 for the implicit match at bob. The second is a flag that is t when - ;; the match is inside a macro. Point should be moved forward by at least - ;; one token. + ;; the match is inside a macro. The third is a flag that is t when the + ;; match is at "top level", i.e. outside any brace block, or directly inside + ;; a class or namespace, etc. Point should be moved forward by at least one + ;; token. ;; ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot, ;; it should return non-nil to ensure that the next search will find them. ＠＠ -6059,6 +6061,8 ＠＠ (setq cfd-macro-end 0) nil)))) ; end of when condition + (when (> cfd-macro-end 0) + (setq cfd-top-level nil)) ; In a macro is "never" at top level. (c-debug-put-decl-spot-faces cfd-match-pos (point)) (if (funcall cfd-fun cfd-match-pos (/= cfd-macro-end 0) cfd-top-level) (setq cfd-prop-match nil)) ＠＠ -8586,7 +8590,13 ＠＠ (looking-at c-noise-macro-with-parens-name-re)) (c-forward-noise-clause)) - ((looking-at c-type-decl-suffix-key) + ((and (looking-at c-type-decl-suffix-key) + ;; We avoid recognizing foo(bar) or foo() at top level as a + ;; construct here in C, since we want to recognize this as a + ;; typeless function declaration. + (not (and (c-major-mode-is 'c-mode) + (eq context 'top) + (eq (char-after) ?\))))) (if (eq (char-after) ?\)) (when (> paren-depth 0) (setq paren-depth (1- paren-depth)) ＠＠ -8629,7 +8639,12 ＠＠ (save-excursion (goto-char after-paren-pos) (c-forward-syntactic-ws) - (c-forward-type))))) + (or (c-forward-type) + ;; Recognize a top-level typeless + ;; function declaration in C. + (and (c-major-mode-is 'c-mode) + (eq context 'top) + (eq (char-after) ?\)))))))) (setq pos (c-up-list-forward (point))) (eq (char-before pos) ?\))) (c-fdoc-shift-type-backward) ＠＠ -9046,9 +9061,12 ＠＠ ;; (in at least C++) that anything that can be parsed as a declaration ;; is a declaration. Now we're being more defensive and prefer to ;; highlight things like "foo (bar);" as a declaration only if we're - ;; inside an arglist that contains declarations. + ;; inside an arglist that contains declarations. Update (2017-09): We + ;; now recognize a top-level "foo(bar);" as a declaration in C. ;; CASE 19 - (eq context 'decl)))) + (or (eq context 'decl) + (and (c-major-mode-is 'c-mode) + (eq context 'top)))))) ;; The point is now after the type decl expression. ＠＠ -9556,6 +9574,7 ＠＠ comment at the start of cc-engine.el for more info." ;; Note to maintainers: this function consumes a great mass of CPU cycles. ;; It's use should thus be minimised as far as possible. + ;; Consider instead using `c-bs-at-toplevel-p'. (let ((paren-state (c-parse-state))) (or (not (c-most-enclosing-brace paren-state)) (c-search-uplist-for-classkey paren-state)))) ＠＠ -9585,8 +9604,15 ＠＠ (not (and (c-major-mode-is 'objc-mode) (c-forward-objc-directive))) + ;; Don't confuse #if .... defined(foo) for a function arglist. + (not (and (looking-at c-cpp-expr-functions-key) + (save-excursion + (save-restriction + (widen) + (c-beginning-of-macro lim))))) (setq id-start (car-safe (c-forward-decl-or-cast-1 (c-point 'bosws) 'top nil))) + (numberp id-start) (< id-start beg) ;; There should not be a '=' or ',' between beg and the diff -r e39b785df77e0a0f89c55aa73e042cc6cc499028 -r 57b32f709a4f3f80c78071d67fec82b4cc629878 cc-langs.el --- a/cc-langs.el +++ b/cc-langs.el ＠＠ -935,6 +935,11 ＠＠ '("defined")) pike '("defined" "efun" "constant")) +(c-lang-defconst c-cpp-expr-functions-key + ;; Matches a function in a cpp expression. + t (c-make-keywords-re t (c-lang-const c-cpp-expr-functions))) +(c-lang-defvar c-cpp-expr-functions-key (c-lang-const c-cpp-expr-functions-key)) + (c-lang-defconst c-assignment-operators "List of all assignment operators." t '("=" "*=" "/=" "%=" "+=" "-=" ">>=" "<<=" "&=" "^=" "|=") diff -r e39b785df77e0a0f89c55aa73e042cc6cc499028 -r 57b32f709a4f3f80c78071d67fec82b4cc629878 cc-mode.el --- a/cc-mode.el +++ b/cc-mode.el ＠＠ -1551,6 +1551,8 ＠＠ (and (c-beginning-of-macro) (progn (c-end-of-macro) (point)))))) (when (and (c-forward-declarator lim) + (or (not (eq (char-after) ?\()) + (c-go-list-forward nil lim)) (eq (c-forward-token-2 1 nil lim) 0)) (c-backward-syntactic-ws) (point)))))) ＠＠ -1569,7 +1571,7 ＠＠ (or (c-fl-decl-start c-new-BEG) (c-point 'bol c-new-BEG)) c-new-END (or (c-fl-decl-end c-new-END) - (c-point 'bonl (max (1- c-new-END) (point-min))))))) + (c-point 'bonl c-new-END))))) (defun c-context-expand-fl-region (beg end) ;; Return a cons (NEW-BEG . NEW-END), where NEW-BEG is the beginning of a Repository URL: https://bitbucket.org/xemacs/cc-mode/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

commit/cc-mode: acm: Fix a CC Mode brace stack cache bug. 7 years, 3 months

Bitbucket

1 new commit in cc-mode: https://bitbucket.org/xemacs/cc-mode/commits/e39b785df77e/ Changeset: e39b785df77e User: acm Date: 2017-10-02 14:21:41+00:00 Summary: Fix a CC Mode brace stack cache bug. * cc-engine.el (c-update-brace-stack): Call c-beginning-of-current-token after a failing search operation, to ensure we don't cache a point inside a token. Affected #: 1 file diff -r 31e4abcac39929366dbd6be1d224351064226973 -r e39b785df77e0a0f89c55aa73e042cc6cc499028 cc-engine.el --- a/cc-engine.el +++ b/cc-engine.el ＠＠ -5420,8 +5420,7 ＠＠ ;; value. (let (match kwd-sym (prev-match-pos 1) (s (cdr stack)) - (bound-<> (car stack)) - ) + (bound-<> (car stack))) (save-excursion (cond ((and bound-<> (<= to bound-<>)) ＠＠ -5482,6 +5481,9 ＠＠ (setq s (cdr s)))) ((c-keyword-member kwd-sym 'c-flat-decl-block-kwds) (push 0 s)))) + ;; The failing `c-syntactic-re-search-forward' may have left us in the + ;; middle of a token, which might be a significant token. Fix this! + (c-beginning-of-current-token) (cons (point) (cons bound-<> s))))) Repository URL: https://bitbucket.org/xemacs/cc-mode/ -- This is a commit notification from bitbucket.org. You are receiving this because you have the service enabled, addressing the recipient of this email.

1 participants
0 comments

2025

2024

2023

2022

2021

2020

2019

2018

2017

XEmacs-commits October 2017