Mercurial > hg > xemacs-beta
view src/lread.c @ 373:6240c7796c7a r21-2b2
Import from CVS: tag r21-2b2
author | cvs |
---|---|
date | Mon, 13 Aug 2007 11:04:06 +0200 |
parents | cc15677e0335 |
children | 8626e4521993 |
line wrap: on
line source
/* Lisp parsing and input streams. Copyright (C) 1985-1989, 1992-1995 Free Software Foundation, Inc. Copyright (C) 1995 Tinker Systems. Copyright (C) 1996 Ben Wing. This file is part of XEmacs. XEmacs is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. XEmacs is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with XEmacs; see the file COPYING. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ /* Synched up with: Mule 2.0, FSF 19.30. */ /* This file has been Mule-ized. */ #include <config.h> #include "lisp.h" #include "buffer.h" #include "bytecode.h" #include "commands.h" #include "insdel.h" #include "lstream.h" #include "opaque.h" #include <paths.h> #ifdef FILE_CODING #include "file-coding.h" #endif #include "sysfile.h" #ifdef LISP_FLOAT_TYPE #define THIS_FILENAME lread #include "sysfloat.h" #endif /* LISP_FLOAT_TYPE */ Lisp_Object Qread_char, Qstandard_input; Lisp_Object Qvariable_documentation; #define LISP_BACKQUOTES #ifdef LISP_BACKQUOTES /* Nonzero means inside a new-style backquote with no surrounding parentheses. Fread initializes this to zero, so we need not specbind it or worry about what happens to it when there is an error. XEmacs: Nested backquotes are perfectly legal and fail utterly with this silliness. */ static int new_backquote_flag, old_backquote_flag; Lisp_Object Qbackquote, Qbacktick, Qcomma, Qcomma_at, Qcomma_dot; #endif Lisp_Object Qvariable_domain; /* I18N3 */ Lisp_Object Vvalues, Vstandard_input, Vafter_load_alist; Lisp_Object Qcurrent_load_list; Lisp_Object Qload, Qload_file_name; Lisp_Object Qlocate_file_hash_table; Lisp_Object Qfset; /* See read_escape() for an explanation of this. */ #if 0 int fail_on_bucky_bit_character_escapes; #endif /* This symbol is also used in fns.c */ #define FEATUREP_SYNTAX #ifdef FEATUREP_SYNTAX Lisp_Object Qfeaturep; #endif /* non-zero if inside `load' */ int load_in_progress; /* Whether Fload_internal() should check whether the .el is newer when loading .elc */ int load_warn_when_source_newer; /* Whether Fload_internal() should check whether the .elc doesn't exist */ int load_warn_when_source_only; /* Whether Fload_internal() should ignore .elc files when no suffix is given */ int load_ignore_elc_files; /* Search path for files to be loaded. */ Lisp_Object Vload_path; /* Search path for files when dumping. */ /* Lisp_Object Vdump_load_path; */ /* This is the user-visible association list that maps features to lists of defs in their load files. */ Lisp_Object Vload_history; /* This is used to build the load history. */ Lisp_Object Vcurrent_load_list; /* Name of file actually being read by `load'. */ Lisp_Object Vload_file_name; /* Same as Vload_file_name but not Lisp-accessible. This ensures that our #$ checks are reliable. */ Lisp_Object Vload_file_name_internal; Lisp_Object Vload_file_name_internal_the_purecopy; /* Function to use for reading, in `load' and friends. */ Lisp_Object Vload_read_function; /* The association list of objects read with the #n=object form. Each member of the list has the form (n . object), and is used to look up the object for the corresponding #n# construct. It must be set to nil before all top-level calls to read0. */ Lisp_Object read_objects; /* Nonzero means load should forcibly load all dynamic doc strings. */ /* Note that this always happens (with some special behavior) when purify_flag is set. */ static int load_force_doc_strings; /* List of descriptors now open for Fload_internal. */ static Lisp_Object Vload_descriptor_list; /* In order to implement "load_force_doc_strings", we keep a list of all the compiled-function objects and such that we have created in the process of loading this file. See the rant below. We specbind this just like Vload_file_name, so there's no problems with recursive loading. */ static Lisp_Object Vload_force_doc_string_list; /* A resizing-buffer stream used to temporarily hold data while reading */ static Lisp_Object Vread_buffer_stream; #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Lisp_Object Vcurrent_compiled_function_annotation; #endif static int load_byte_code_version; /* An array describing all known built-in structure types */ static structure_type_dynarr *the_structure_type_dynarr; #if 0 /* FSF defun hack */ /* When nonzero, read conses in pure space */ static int read_pure; #endif #if 0 /* FSF stuff */ /* For use within read-from-string (this reader is non-reentrant!!) */ static int read_from_string_index; static int read_from_string_limit; #endif #if 0 /* More FSF implementation kludges. */ /* In order to implement load-force-doc-string, FSF saves the #@-quoted string when it's seen, and goes back and retrieves it later. This approach is not only kludgy, but it in general won't work correctly because there's no stack of remembered #@-quoted-strings and those strings don't generally appear in the file in the same order as their #$ references. (Yes, that is amazingly stupid too. It would be trivially easy to always encode the #@ string [which is a comment, anyway] in the middle of the (#$ . INT) cons reference. That way, it would be really easy to implement load-force-doc-string in a non-kludgy way by just retrieving the string immediately, because it's delivered on a silver platter.) And finally, this stupid approach doesn't work under Mule, or under MS-DOS or Windows NT, or under VMS, or any other place where you either can't do an ftell() or don't get back a byte count. Oh, and one more lossage in this approach: If you attempt to dump any ELC files that were compiled with `byte-compile-dynamic' (as opposed to just `byte-compile-dynamic-docstring'), you get hosed. FMH! (as the illustrious JWZ was prone to utter) The approach we use is clean, solves all of these problems, and is probably easier to implement anyway. We just save a list of all the containing objects that have (#$ . INT) conses in them (this will only be compiled-function objects and lists), and when the file is finished loading, we go through and fill in all the doc strings at once. */ /* This contains the last string skipped with #@. */ static char *saved_doc_string; /* Length of buffer allocated in saved_doc_string. */ static int saved_doc_string_size; /* Length of actual data in saved_doc_string. */ static int saved_doc_string_length; /* This is the file position that string came from. */ static int saved_doc_string_position; #endif EXFUN (Fread_from_string, 3); /* When errors are signaled, the actual readcharfun should not be used as an argument if it is an lstream, so that lstreams don't escape to the Lisp level. */ #define READCHARFUN_MAYBE(x) (LSTREAMP (x) \ ? (build_string ("internal input stream")) \ : (x)) static DOESNT_RETURN syntax_error (CONST char *string) { signal_error (Qinvalid_read_syntax, list1 (build_translated_string (string))); } static Lisp_Object continuable_syntax_error (CONST char *string) { return Fsignal (Qinvalid_read_syntax, list1 (build_translated_string (string))); } /* Handle unreading and rereading of characters. */ static Emchar readchar (Lisp_Object readcharfun) { /* This function can GC */ if (BUFFERP (readcharfun)) { Emchar c; struct buffer *b = XBUFFER (readcharfun); if (!BUFFER_LIVE_P (b)) error ("Reading from killed buffer"); if (BUF_PT (b) >= BUF_ZV (b)) return -1; c = BUF_FETCH_CHAR (b, BUF_PT (b)); BUF_SET_PT (b, BUF_PT (b) + 1); return c; } else if (LSTREAMP (readcharfun)) { Emchar c = Lstream_get_emchar (XLSTREAM (readcharfun)); #ifdef DEBUG_XEMACS /* testing Mule */ static int testing_mule = 0; /* Change via debugger */ if (testing_mule) { if (c >= 0x20 && c <= 0x7E) fprintf (stderr, "%c", c); else if (c == '\n') fprintf (stderr, "\\n\n"); else fprintf (stderr, "\\%o ", c); } #endif return c; } else if (MARKERP (readcharfun)) { Emchar c; Bufpos mpos = marker_position (readcharfun); struct buffer *inbuffer = XMARKER (readcharfun)->buffer; if (mpos >= BUF_ZV (inbuffer)) return -1; c = BUF_FETCH_CHAR (inbuffer, mpos); set_marker_position (readcharfun, mpos + 1); return c; } else { Lisp_Object tem = call0 (readcharfun); if (!CHAR_OR_CHAR_INTP (tem)) return -1; return XCHAR_OR_CHAR_INT (tem); } } /* Unread the character C in the way appropriate for the stream READCHARFUN. If the stream is a user function, call it with the char as argument. */ static void unreadchar (Lisp_Object readcharfun, Emchar c) { if (c == -1) /* Don't back up the pointer if we're unreading the end-of-input mark, since readchar didn't advance it when we read it. */ ; else if (BUFFERP (readcharfun)) BUF_SET_PT (XBUFFER (readcharfun), BUF_PT (XBUFFER (readcharfun)) - 1); else if (LSTREAMP (readcharfun)) { Lstream_unget_emchar (XLSTREAM (readcharfun), c); #ifdef DEBUG_XEMACS /* testing Mule */ { static int testing_mule = 0; /* Set this using debugger */ if (testing_mule) fprintf (stderr, (c >= 0x20 && c <= 0x7E) ? "UU%c" : ((c == '\n') ? "UU\\n\n" : "UU\\%o"), c); } #endif } else if (MARKERP (readcharfun)) set_marker_position (readcharfun, marker_position (readcharfun) - 1); else call1 (readcharfun, make_char (c)); } static Lisp_Object read0 (Lisp_Object readcharfun); static Lisp_Object read1 (Lisp_Object readcharfun); /* allow_dotted_lists means that something like (foo bar . baz) is acceptable. If -1, means check for starting with defun and make structure pure. (not implemented, probably for very good reasons) */ /* If check_for_doc_references, look for (#$ . INT) doc references in the list and record if load_force_doc_strings is non-zero. (Such doc references will be destroyed during the loadup phase by replacing with Qzero, because Snarf-documentation will fill them in again.) WARNING: If you set this, you sure as hell better not call free_list() on the returned list here. */ static Lisp_Object read_list (Lisp_Object readcharfun, Emchar terminator, int allow_dotted_lists, int check_for_doc_references); static void readevalloop (Lisp_Object readcharfun, Lisp_Object sourcefile, Lisp_Object (*evalfun) (Lisp_Object), int printflag); static Lisp_Object load_unwind (Lisp_Object stream) /* used as unwind-protect function in load */ { Lstream_close (XLSTREAM (stream)); if (--load_in_progress < 0) load_in_progress = 0; return Qnil; } static Lisp_Object load_descriptor_unwind (Lisp_Object oldlist) { Vload_descriptor_list = oldlist; return Qnil; } static Lisp_Object load_file_name_internal_unwind (Lisp_Object oldval) { Vload_file_name_internal = oldval; return Qnil; } static Lisp_Object load_file_name_internal_the_purecopy_unwind (Lisp_Object oldval) { Vload_file_name_internal_the_purecopy = oldval; return Qnil; } static Lisp_Object load_byte_code_version_unwind (Lisp_Object oldval) { load_byte_code_version = XINT (oldval); return Qnil; } /* The plague is coming. Ring around the rosy, pocket full of posy, Ashes ashes, they all fall down. */ void ebolify_bytecode_constants (Lisp_Object vector) { int len = XVECTOR_LENGTH (vector); int i; for (i = 0; i < len; i++) { Lisp_Object el = XVECTOR_DATA (vector)[i]; /* We don't check for `eq', `equal', and the others that have bytecode opcodes. This might lose if someone passes #'eq or something to `funcall', but who would really do that? As they say in law, we've made a "good-faith effort" to unfuckify ourselves. And doing it this way avoids screwing up args to `make-hashtable' and such. As it is, we have to add an extra Ebola check in decode_weak_list_type(). --ben */ if (EQ (el, Qassoc)) el = Qold_assoc; if (EQ (el, Qdelq)) el = Qold_delq; #if 0 /* I think this is a bad idea because it will probably mess with keymap code. */ if (EQ (el, Qdelete)) el = Qold_delete; #endif if (EQ (el, Qrassq)) el = Qold_rassq; if (EQ (el, Qrassoc)) el = Qold_rassoc; XVECTOR_DATA (vector)[i] = el; } } static Lisp_Object pas_de_lache_ici (int fd, Lisp_Object victim) { Lisp_Object tem; EMACS_INT pos; if (!INTP (XCDR (victim))) signal_simple_error ("Bogus doc string reference", victim); pos = XINT (XCDR (victim)); if (pos < 0) pos = -pos; /* kludge to mark a user variable */ tem = unparesseuxify_doc_string (fd, pos, 0, Vload_file_name_internal); if (!STRINGP (tem)) signal_error (Qerror, tem); return tem; } static Lisp_Object load_force_doc_string_unwind (Lisp_Object oldlist) { struct gcpro gcpro1; Lisp_Object list = Vload_force_doc_string_list; Lisp_Object tail; int fd = XINT (XCAR (Vload_descriptor_list)); /* NOTE: If purify_flag is true, we're in-place modifying objects that may be in purespace (and if not, they will be). Therefore, we have to be VERY careful to make sure that all objects that we create are purecopied -- objects in purespace are not marked for GC, and if we leave any impure objects inside of pure ones, we're really screwed. */ GCPRO1 (list); /* restore the old value first just in case an error occurs. */ Vload_force_doc_string_list = oldlist; LIST_LOOP (tail, list) { Lisp_Object john = Fcar (tail); if (CONSP (john)) { assert (CONSP (XCAR (john))); assert (!purify_flag); /* should have been handled in read_list() */ XCAR (john) = pas_de_lache_ici (fd, XCAR (john)); } else { Lisp_Object doc; assert (COMPILED_FUNCTIONP (john)); if (CONSP (XCOMPILED_FUNCTION (john)->bytecodes)) { struct gcpro ngcpro1; Lisp_Object juan = (pas_de_lache_ici (fd, XCOMPILED_FUNCTION (john)->bytecodes)); Lisp_Object ivan; NGCPRO1 (juan); ivan = Fread (juan); if (!CONSP (ivan)) signal_simple_error ("invalid lazy-loaded byte code", ivan); /* Remember to purecopy; see above. */ XCOMPILED_FUNCTION (john)->bytecodes = Fpurecopy (XCAR (ivan)); /* v18 or v19 bytecode file. Need to Ebolify. */ if (XCOMPILED_FUNCTION (john)->flags.ebolified && VECTORP (XCDR (ivan))) ebolify_bytecode_constants (XCDR (ivan)); XCOMPILED_FUNCTION (john)->constants = Fpurecopy (XCDR (ivan)); NUNGCPRO; } doc = compiled_function_documentation (XCOMPILED_FUNCTION (john)); if (CONSP (doc)) { assert (!purify_flag); /* should have been handled in read_compiled_function() */ doc = pas_de_lache_ici (fd, doc); set_compiled_function_documentation (XCOMPILED_FUNCTION (john), doc); } } } if (!NILP (list)) free_list (list); UNGCPRO; return Qnil; } /* Close all descriptors in use for Fload_internal. This is used when starting a subprocess. */ void close_load_descs (void) { Lisp_Object tail; LIST_LOOP (tail, Vload_descriptor_list) close (XINT (XCAR (tail))); } #ifdef I18N3 Lisp_Object Vfile_domain; Lisp_Object restore_file_domain (Lisp_Object val) { Vfile_domain = val; return Qnil; } #endif /* I18N3 */ DEFUN ("load-internal", Fload_internal, 1, 6, 0, /* Execute a file of Lisp code named FILE; no coding-system frobbing. This function is identical to `load' except for the handling of the CODESYS and USED-CODESYS arguments under XEmacs/Mule. (When Mule support is not present, both functions are identical and ignore the CODESYS and USED-CODESYS arguments.) If support for Mule exists in this Emacs, the file is decoded according to CODESYS; if omitted, no conversion happens. If USED-CODESYS is non-nil, it should be a symbol, and the actual coding system that was used for the decoding is stored into it. It will in general be different from CODESYS if CODESYS specifies automatic encoding detection or end-of-line detection. */ (file, no_error, nomessage, nosuffix, codesys, used_codesys)) { /* This function can GC */ int fd = -1; int speccount = specpdl_depth (); int source_only = 0; Lisp_Object newer = Qnil; Lisp_Object handler = Qnil; Lisp_Object found = Qnil; struct gcpro gcpro1, gcpro2, gcpro3; int reading_elc = 0; int message_p = NILP (nomessage); /*#ifdef DEBUG_XEMACS*/ static Lisp_Object last_file_loaded; size_t pure_usage = 0; /*#endif*/ struct stat s1, s2; GCPRO3 (file, newer, found); CHECK_STRING (file); /*#ifdef DEBUG_XEMACS*/ if (purify_flag && noninteractive) { message_p = 1; last_file_loaded = file; pure_usage = purespace_usage (); } /*#endif / * DEBUG_XEMACS */ /* If file name is magic, call the handler. */ handler = Ffind_file_name_handler (file, Qload); if (!NILP (handler)) RETURN_UNGCPRO (call5 (handler, Qload, file, no_error, nomessage, nosuffix)); /* Do this after the handler to avoid the need to gcpro noerror, nomessage and nosuffix. (Below here, we care only whether they are nil or not.) */ file = Fsubstitute_in_file_name (file); #ifdef FILE_CODING if (!NILP (used_codesys)) CHECK_SYMBOL (used_codesys); #endif /* Avoid weird lossage with null string as arg, since it would try to load a directory as a Lisp file. Unix truly sucks. */ if (XSTRING_LENGTH (file) > 0) { char *foundstr; int foundlen; fd = locate_file (Vload_path, file, ((!NILP (nosuffix)) ? "" : load_ignore_elc_files ? ".el:" : ".elc:.el:"), &found, -1); if (fd < 0) { if (NILP (no_error)) signal_file_error ("Cannot open load file", file); else { UNGCPRO; return Qnil; } } foundstr = (char *) alloca (XSTRING_LENGTH (found) + 1); strcpy (foundstr, (char *) XSTRING_DATA (found)); foundlen = strlen (foundstr); /* The omniscient JWZ thinks this is worthless, but I beg to differ. --ben */ if (load_ignore_elc_files) { newer = Ffile_name_nondirectory (found); } else if (load_warn_when_source_newer && !memcmp (".elc", foundstr + foundlen - 4, 4)) { if (! fstat (fd, &s1)) /* can't fail, right? */ { int result; /* temporarily hack the 'c' off the end of the filename */ foundstr[foundlen - 1] = '\0'; result = stat (foundstr, &s2); if (result >= 0 && (unsigned) s1.st_mtime < (unsigned) s2.st_mtime) { Lisp_Object newer_name = make_string ((Bufbyte *) foundstr, foundlen - 1); struct gcpro nngcpro1; NNGCPRO1 (newer_name); newer = Ffile_name_nondirectory (newer_name); NNUNGCPRO; } /* put the 'c' back on (kludge-o-rama) */ foundstr[foundlen - 1] = 'c'; } } else if (load_warn_when_source_only && /* `found' ends in ".el" */ !memcmp (".el", foundstr + foundlen - 3, 3) && /* `file' does not end in ".el" */ memcmp (".el", XSTRING_DATA (file) + XSTRING_LENGTH (file) - 3, 3)) { source_only = 1; } if (!memcmp (".elc", foundstr + foundlen - 4, 4)) reading_elc = 1; } #define PRINT_LOADING_MESSAGE(done) do { \ if (load_ignore_elc_files) \ { \ if (message_p) \ message ("Loading %s..." done, XSTRING_DATA (newer)); \ } \ else if (!NILP (newer)) \ message ("Loading %s..." done " (file %s is newer)", \ XSTRING_DATA (file), \ XSTRING_DATA (newer)); \ else if (source_only) \ message ("Loading %s..." done " (file %s.elc does not exist)", \ XSTRING_DATA (file), \ XSTRING_DATA (Ffile_name_nondirectory (file))); \ else if (message_p) \ message ("Loading %s..." done, XSTRING_DATA (file)); \ } while (0) PRINT_LOADING_MESSAGE (""); { /* Lisp_Object's must be malloc'ed, not stack-allocated */ Lisp_Object lispstream = Qnil; CONST int block_size = 8192; struct gcpro ngcpro1; NGCPRO1 (lispstream); lispstream = make_filedesc_input_stream (fd, 0, -1, LSTR_CLOSING); /* 64K is used for normal files; 8K should be OK here because Lisp files aren't really all that big. */ Lstream_set_buffering (XLSTREAM (lispstream), LSTREAM_BLOCKN_BUFFERED, block_size); #ifdef FILE_CODING lispstream = make_decoding_input_stream (XLSTREAM (lispstream), Fget_coding_system (codesys)); Lstream_set_buffering (XLSTREAM (lispstream), LSTREAM_BLOCKN_BUFFERED, block_size); #endif /* NOTE: Order of these is very important. Don't rearrange them. */ record_unwind_protect (load_unwind, lispstream); record_unwind_protect (load_descriptor_unwind, Vload_descriptor_list); record_unwind_protect (load_file_name_internal_unwind, Vload_file_name_internal); record_unwind_protect (load_file_name_internal_the_purecopy_unwind, Vload_file_name_internal_the_purecopy); record_unwind_protect (load_force_doc_string_unwind, Vload_force_doc_string_list); Vload_file_name_internal = found; Vload_file_name_internal_the_purecopy = Qnil; specbind (Qload_file_name, found); Vload_descriptor_list = Fcons (make_int (fd), Vload_descriptor_list); Vload_force_doc_string_list = Qnil; #ifdef I18N3 record_unwind_protect (restore_file_domain, Vfile_domain); Vfile_domain = Qnil; /* set it to nil; a call to #'domain will set it. */ #endif load_in_progress++; /* Now determine what sort of ELC file we're reading in. */ record_unwind_protect (load_byte_code_version_unwind, make_int (load_byte_code_version)); if (reading_elc) { char elc_header[8]; int num_read; num_read = Lstream_read (XLSTREAM (lispstream), elc_header, 8); if (num_read < 8 || strncmp (elc_header, ";ELC", 4)) { /* Huh? Probably not a valid ELC file. */ load_byte_code_version = 100; /* no Ebolification needed */ Lstream_unread (XLSTREAM (lispstream), elc_header, num_read); } else load_byte_code_version = elc_header[4]; } else load_byte_code_version = 100; /* no Ebolification needed */ readevalloop (lispstream, file, Feval, 0); #ifdef FILE_CODING if (!NILP (used_codesys)) Fset (used_codesys, XCODING_SYSTEM_NAME (decoding_stream_coding_system (XLSTREAM (lispstream)))); #endif unbind_to (speccount, Qnil); NUNGCPRO; } { Lisp_Object tem; /* #### Disgusting kludge */ /* Run any load-hooks for this file. */ /* #### An even more disgusting kludge. There is horrible code */ /* that is relying on the fact that dumped lisp files are found */ /* via `load-path' search. */ Lisp_Object name = file; if (!NILP(Ffile_name_absolute_p(file))) { name = Ffile_name_nondirectory(file); } { struct gcpro ngcpro1; NGCPRO1 (name); tem = Fassoc (name, Vafter_load_alist); NUNGCPRO; } if (!NILP (tem)) { struct gcpro ngcpro1; NGCPRO1 (tem); /* Use eval so that errors give a semi-meaningful backtrace. --Stig */ tem = Fcons (Qprogn, Fcdr (tem)); Feval (tem); NUNGCPRO; } } /*#ifdef DEBUG_XEMACS*/ if (purify_flag && noninteractive) { if (EQ (last_file_loaded, file)) message_append (" (%ld)", (unsigned long) (purespace_usage() - pure_usage)); else message ("Loading %s ...done (%ld)", XSTRING_DATA (file), (unsigned long) (purespace_usage() - pure_usage)); } /*#endif / * DEBUG_XEMACS */ if (!noninteractive) PRINT_LOADING_MESSAGE ("done"); UNGCPRO; return Qt; } #if 0 /* FSFmacs */ /* not used */ static int complete_filename_p (Lisp_Object pathname) { REGISTER unsigned char *s = XSTRING_DATA (pathname); return (IS_DIRECTORY_SEP (s[0]) || (XSTRING_LENGTH (pathname) > 2 && IS_DEVICE_SEP (s[1]) && IS_DIRECTORY_SEP (s[2])) #ifdef ALTOS || *s == '@' #endif ); } #endif /* 0 */ DEFUN ("locate-file", Flocate_file, 2, 4, 0, /* Search for FILENAME through PATH-LIST, expanded by one of the optional SUFFIXES (string of suffixes separated by ":"s), checking for access MODE (0|1|2|4 = exists|executable|writeable|readable), default readable. `locate-file' keeps hash tables of the directories it searches through, in order to speed things up. It tries valiantly to not get confused in the face of a changing and unpredictable environment, but can occasionally get tripped up. In this case, you will have to call `locate-file-clear-hashing' to get it back on track. See that function for details. */ (filename, path_list, suffixes, mode)) { /* This function can GC */ Lisp_Object tp; CHECK_STRING (filename); if (!NILP (suffixes)) CHECK_STRING (suffixes); if (!NILP (mode)) CHECK_NATNUM (mode); locate_file (path_list, filename, ((NILP (suffixes)) ? "" : (char *) (XSTRING_DATA (suffixes))), &tp, (NILP (mode) ? R_OK : XINT (mode))); return tp; } /* recalculate the hash table for the given string */ static Lisp_Object locate_file_refresh_hashing (Lisp_Object str) { Lisp_Object hash = make_directory_hash_table ((char *) XSTRING_DATA (str)); Fput (str, Qlocate_file_hash_table, hash); return hash; } /* find the hash table for the given string, recalculating if necessary */ static Lisp_Object locate_file_find_directory_hash_table (Lisp_Object str) { Lisp_Object hash = Fget (str, Qlocate_file_hash_table, Qnil); if (NILP (Fhashtablep (hash))) return locate_file_refresh_hashing (str); return hash; } /* look for STR in PATH, optionally adding suffixes in SUFFIX */ static int locate_file_in_directory (Lisp_Object path, Lisp_Object str, CONST char *suffix, Lisp_Object *storeptr, int mode) { /* This function can GC */ int fd; int fn_size = 100; char buf[100]; char *fn = buf; int want_size; struct stat st; Lisp_Object filename = Qnil; struct gcpro gcpro1, gcpro2, gcpro3; CONST char *nsuffix; GCPRO3 (path, str, filename); filename = Fexpand_file_name (str, path); if (NILP (filename) || NILP (Ffile_name_absolute_p (filename))) /* If there are non-absolute elts in PATH (eg ".") */ /* Of course, this could conceivably lose if luser sets default-directory to be something non-absolute ... */ { if (NILP (filename)) /* NIL means current dirctory */ filename = current_buffer->directory; else filename = Fexpand_file_name (filename, current_buffer->directory); if (NILP (Ffile_name_absolute_p (filename))) { /* Give up on this path element! */ UNGCPRO; return -1; } } /* Calculate maximum size of any filename made from this path element/specified file name and any possible suffix. */ want_size = strlen (suffix) + XSTRING_LENGTH (filename) + 1; if (fn_size < want_size) fn = (char *) alloca (fn_size = 100 + want_size); nsuffix = suffix; /* Loop over suffixes. */ while (1) { char *esuffix = (char *) strchr (nsuffix, ':'); int lsuffix = ((esuffix) ? (esuffix - nsuffix) : strlen (nsuffix)); /* Concatenate path element/specified name with the suffix. */ strncpy (fn, (char *) XSTRING_DATA (filename), XSTRING_LENGTH (filename)); fn[XSTRING_LENGTH (filename)] = 0; if (lsuffix != 0) /* Bug happens on CCI if lsuffix is 0. */ strncat (fn, nsuffix, lsuffix); /* Ignore file if it's a directory. */ if (stat (fn, &st) >= 0 && (st.st_mode & S_IFMT) != S_IFDIR) { /* Check that we can access or open it. */ if (mode >= 0) fd = access (fn, mode); else fd = open (fn, O_RDONLY | OPEN_BINARY, 0); if (fd >= 0) { /* We succeeded; return this descriptor and filename. */ if (storeptr) *storeptr = build_string (fn); UNGCPRO; #ifndef WINDOWSNT /* If we actually opened the file, set close-on-exec flag on the new descriptor so that subprocesses can't whack at it. */ if (mode < 0) (void) fcntl (fd, F_SETFD, FD_CLOEXEC); #endif return fd; } } /* Advance to next suffix. */ if (esuffix == 0) break; nsuffix += lsuffix + 1; } UNGCPRO; return -1; } /* do the same as locate_file() but don't use any hash tables. */ static int locate_file_without_hash (Lisp_Object path, Lisp_Object str, CONST char *suffix, Lisp_Object *storeptr, int mode) { /* This function can GC */ int absolute; struct gcpro gcpro1; /* is this necessary? */ GCPRO1 (path); absolute = !NILP (Ffile_name_absolute_p (str)); for (; !NILP (path); path = Fcdr (path)) { int val = locate_file_in_directory (Fcar (path), str, suffix, storeptr, mode); if (val >= 0) { UNGCPRO; return val; } if (absolute) break; } UNGCPRO; return -1; } /* Construct a list of all files to search for. */ static Lisp_Object locate_file_construct_suffixed_files (Lisp_Object str, CONST char *suffix) { int want_size; int fn_size = 100; char buf[100]; char *fn = buf; CONST char *nsuffix; Lisp_Object suffixtab = Qnil; /* Calculate maximum size of any filename made from this path element/specified file name and any possible suffix. */ want_size = strlen (suffix) + XSTRING_LENGTH (str) + 1; if (fn_size < want_size) fn = (char *) alloca (fn_size = 100 + want_size); nsuffix = suffix; while (1) { char *esuffix = (char *) strchr (nsuffix, ':'); int lsuffix = ((esuffix) ? (esuffix - nsuffix) : strlen (nsuffix)); /* Concatenate path element/specified name with the suffix. */ strncpy (fn, (char *) XSTRING_DATA (str), XSTRING_LENGTH (str)); fn[XSTRING_LENGTH (str)] = 0; if (lsuffix != 0) /* Bug happens on CCI if lsuffix is 0. */ strncat (fn, nsuffix, lsuffix); suffixtab = Fcons (build_string (fn), suffixtab); /* Advance to next suffix. */ if (esuffix == 0) break; nsuffix += lsuffix + 1; } return Fnreverse (suffixtab); } DEFUN ("locate-file-clear-hashing", Flocate_file_clear_hashing, 1, 1, 0, /* Clear the hash records for the specified list of directories. `locate-file' uses a hashing scheme to speed lookup, and will correctly track the following environmental changes: -- changes of any sort to the list of directories to be searched. -- addition and deletion of non-shadowing files (see below) from the directories in the list. -- byte-compilation of a .el file into a .elc file. `locate-file' will primarily get confused if you add a file that shadows \(i.e. has the same name as) another file further down in the directory list. In this case, you must call `locate-file-clear-hashing'. */ (path)) { Lisp_Object pathtail; for (pathtail = path; !NILP (pathtail); pathtail = Fcdr (pathtail)) { Lisp_Object pathel = Fcar (pathtail); if (!purified (pathel)) Fput (pathel, Qlocate_file_hash_table, Qnil); } return Qnil; } /* Search for a file whose name is STR, looking in directories in the Lisp list PATH, and trying suffixes from SUFFIX. SUFFIX is a string containing possible suffixes separated by colons. On success, returns a file descriptor. On failure, returns -1. MODE nonnegative means don't open the files, just look for one for which access(file,MODE) succeeds. In this case, returns 1 on success. If STOREPTR is nonzero, it points to a slot where the name of the file actually found should be stored as a Lisp string. Nil is stored there on failure. Called openp() in FSFmacs. */ int locate_file (Lisp_Object path, Lisp_Object str, CONST char *suffix, Lisp_Object *storeptr, int mode) { /* This function can GC */ Lisp_Object suffixtab = Qnil; Lisp_Object pathtail; int val; struct gcpro gcpro1, gcpro2, gcpro3; if (storeptr) *storeptr = Qnil; /* if this filename has directory components, it's too complicated to try and use the hash tables. */ if (!NILP (Ffile_name_directory (str))) return locate_file_without_hash (path, str, suffix, storeptr, mode); /* Is it really necessary to gcpro path and str? It shouldn't be unless some caller has fucked up. */ GCPRO3 (path, str, suffixtab); suffixtab = locate_file_construct_suffixed_files (str, suffix); for (pathtail = path; !NILP (pathtail); pathtail = Fcdr (pathtail)) { Lisp_Object pathel = Fcar (pathtail); Lisp_Object hashtab; Lisp_Object tail; int found; /* If this path element is relative, we have to look by hand. Can't set string property in a pure string. */ if (NILP (pathel) || NILP (Ffile_name_absolute_p (pathel)) || purified (pathel)) { val = locate_file_in_directory (pathel, str, suffix, storeptr, mode); if (val >= 0) { UNGCPRO; return val; } continue; } hashtab = locate_file_find_directory_hash_table (pathel); /* Loop over suffixes. */ for (tail = suffixtab, found = 0; !found && CONSP (tail); tail = XCDR (tail)) { if (!NILP (Fgethash (XCAR (tail), hashtab, Qnil))) found = 1; } if (found) { /* This is a likely candidate. Look by hand in this directory so we don't get thrown off if someone byte-compiles a file. */ val = locate_file_in_directory (pathel, str, suffix, storeptr, mode); if (val >= 0) { UNGCPRO; return val; } /* Hmm ... the file isn't actually there. (Or possibly it's a directory ...) So refresh our hashing. */ locate_file_refresh_hashing (pathel); } } /* File is probably not there, but check the hard way just in case. */ val = locate_file_without_hash (path, str, suffix, storeptr, mode); if (val >= 0) { /* Sneaky user added a file without telling us. */ Flocate_file_clear_hashing (path); } UNGCPRO; return val; } #ifdef LOADHIST /* Merge the list we've accumulated of globals from the current input source into the load_history variable. The details depend on whether the source has an associated file name or not. */ static void build_load_history (int loading, Lisp_Object source) { REGISTER Lisp_Object tail, prev, newelt; REGISTER Lisp_Object tem, tem2; int foundit; #if !defined(LOADHIST_DUMPED) /* Don't bother recording anything for preloaded files. */ if (purify_flag) return; #endif tail = Vload_history; prev = Qnil; foundit = 0; while (!NILP (tail)) { tem = Fcar (tail); /* Find the feature's previous assoc list... */ if (internal_equal (source, Fcar (tem), 0)) { foundit = 1; /* If we're loading, remove it. */ if (loading) { if (NILP (prev)) Vload_history = Fcdr (tail); else Fsetcdr (prev, Fcdr (tail)); } /* Otherwise, cons on new symbols that are not already members. */ else { tem2 = Vcurrent_load_list; while (CONSP (tem2)) { newelt = XCAR (tem2); if (NILP (Fmemq (newelt, tem))) Fsetcar (tail, Fcons (Fcar (tem), Fcons (newelt, Fcdr (tem)))); tem2 = XCDR (tem2); QUIT; } } } else prev = tail; tail = Fcdr (tail); QUIT; } /* If we're loading, cons the new assoc onto the front of load-history, the most-recently-loaded position. Also do this if we didn't find an existing member for the current source. */ if (loading || !foundit) Vload_history = Fcons (Fnreverse (Vcurrent_load_list), Vload_history); } #else /* !LOADHIST */ #define build_load_history(x,y) #endif /* !LOADHIST */ #if 0 /* FSFmacs defun hack */ Lisp_Object unreadpure (void) /* Used as unwind-protect function in readevalloop */ { read_pure = 0; return Qnil; } #endif /* 0 */ static void readevalloop (Lisp_Object readcharfun, Lisp_Object sourcename, Lisp_Object (*evalfun) (Lisp_Object), int printflag) { /* This function can GC */ REGISTER Emchar c; REGISTER Lisp_Object val; int speccount = specpdl_depth (); struct gcpro gcpro1; struct buffer *b = 0; if (BUFFERP (readcharfun)) b = XBUFFER (readcharfun); else if (MARKERP (readcharfun)) b = XMARKER (readcharfun)->buffer; /* Don't do this. It is not necessary, and it needlessly exposes READCHARFUN (which can be a stream) to Lisp. --hniksic */ /*specbind (Qstandard_input, readcharfun);*/ specbind (Qcurrent_load_list, Qnil); #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Vcurrent_compiled_function_annotation = Qnil; #endif GCPRO1 (sourcename); LOADHIST_ATTACH (sourcename); while (1) { QUIT; if (b != 0 && !BUFFER_LIVE_P (b)) error ("Reading from killed buffer"); c = readchar (readcharfun); if (c == ';') { /* Skip comment */ while ((c = readchar (readcharfun)) != '\n' && c != -1) QUIT; continue; } if (c < 0) break; /* Ignore whitespace here, so we can detect eof. */ if (c == ' ' || c == '\t' || c == '\n' || c == '\f' || c == '\r') continue; #if 0 /* FSFmacs defun hack */ if (purify_flag && c == '(') { int count1 = specpdl_depth (); record_unwind_protect (unreadpure, Qnil); val = read_list (readcharfun, ')', -1, 1); unbind_to (count1, Qnil); } else #else /* No "defun hack" -- Emacs 19 uses read-time syntax for bytecodes */ { unreadchar (readcharfun, c); read_objects = Qnil; if (NILP (Vload_read_function)) val = read0 (readcharfun); else val = call1 (Vload_read_function, readcharfun); } #endif val = (*evalfun) (val); if (printflag) { Vvalues = Fcons (val, Vvalues); if (EQ (Vstandard_output, Qt)) Fprin1 (val, Qnil); else Fprint (val, Qnil); } } build_load_history (LSTREAMP (readcharfun) || /* This looks weird, but it's what's in FSFmacs */ (b ? BUF_NARROWED (b) : BUF_NARROWED (current_buffer)), sourcename); UNGCPRO; unbind_to (speccount, Qnil); } DEFUN ("eval-buffer", Feval_buffer, 0, 2, "bBuffer: ", /* Execute BUFFER as Lisp code. Programs can pass two arguments, BUFFER and PRINTFLAG. BUFFER is the buffer to evaluate (nil means use current buffer). PRINTFLAG controls printing of output: nil means discard it; anything else is stream for print. If there is no error, point does not move. If there is an error, point remains at the end of the last character read from the buffer. Execute BUFFER as Lisp code. */ (bufname, printflag)) { /* This function can GC */ int speccount = specpdl_depth (); Lisp_Object tem, buf; if (NILP (bufname)) buf = Fcurrent_buffer (); else buf = Fget_buffer (bufname); if (NILP (buf)) error ("No such buffer."); if (NILP (printflag)) tem = Qsymbolp; /* #### #@[]*&$#*[& SI:NULL-STREAM */ else tem = printflag; specbind (Qstandard_output, tem); record_unwind_protect (save_excursion_restore, save_excursion_save ()); BUF_SET_PT (XBUFFER (buf), BUF_BEGV (XBUFFER (buf))); readevalloop (buf, XBUFFER (buf)->filename, Feval, !NILP (printflag)); return unbind_to (speccount, Qnil); } #if 0 xxDEFUN ("eval-current-buffer", Feval_current_buffer, 0, 1, "", /* Execute the current buffer as Lisp code. Programs can pass argument PRINTFLAG which controls printing of output: nil means discard it; anything else is stream for print. If there is no error, point does not move. If there is an error, point remains at the end of the last character read from the buffer. */ (printflag)) { code omitted; } #endif /* 0 */ DEFUN ("eval-region", Feval_region, 2, 3, "r", /* Execute the region as Lisp code. When called from programs, expects two arguments, giving starting and ending indices in the current buffer of the text to be executed. Programs can pass third argument PRINTFLAG which controls output: nil means discard it; anything else is stream for printing it. If there is no error, point does not move. If there is an error, point remains at the end of the last character read from the buffer. Note: Before evaling the region, this function narrows the buffer to it. If the code being eval'd should happen to trigger a redisplay you may see some text temporarily disappear because of this. */ (b, e, printflag)) { /* This function can GC */ int speccount = specpdl_depth (); Lisp_Object tem; Lisp_Object cbuf = Fcurrent_buffer (); if (NILP (printflag)) tem = Qsymbolp; /* #### #@[]*&$#*[& SI:NULL-STREAM */ else tem = printflag; specbind (Qstandard_output, tem); if (NILP (printflag)) record_unwind_protect (save_excursion_restore, save_excursion_save ()); record_unwind_protect (save_restriction_restore, save_restriction_save ()); /* This both uses b and checks its type. */ Fgoto_char (b, cbuf); Fnarrow_to_region (make_int (BUF_BEGV (current_buffer)), e, cbuf); readevalloop (cbuf, XBUFFER (cbuf)->filename, Feval, !NILP (printflag)); return unbind_to (speccount, Qnil); } DEFUN ("read", Fread, 0, 1, 0, /* Read one Lisp expression as text from STREAM, return as Lisp object. If STREAM is nil, use the value of `standard-input' (which see). STREAM or the value of `standard-input' may be: a buffer (read from point and advance it) a marker (read from where it points and advance it) a function (call it with no arguments for each character, call it with a char as argument to push a char back) a string (takes text from string, starting at the beginning) t (read text line using minibuffer and use it). */ (stream)) { if (NILP (stream)) stream = Vstandard_input; if (EQ (stream, Qt)) stream = Qread_char; read_objects = Qnil; #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Vcurrent_compiled_function_annotation = Qnil; #endif if (EQ (stream, Qread_char)) { Lisp_Object val = call1 (Qread_from_minibuffer, build_translated_string ("Lisp expression: ")); return Fcar (Fread_from_string (val, Qnil, Qnil)); } if (STRINGP (stream)) return Fcar (Fread_from_string (stream, Qnil, Qnil)); return read0 (stream); } DEFUN ("read-from-string", Fread_from_string, 1, 3, 0, /* Read one Lisp expression which is represented as text by STRING. Returns a cons: (OBJECT-READ . FINAL-STRING-INDEX). START and END optionally delimit a substring of STRING from which to read; they default to 0 and (length STRING) respectively. */ (string, start, end)) { Bytecount startval, endval; Lisp_Object tem; Lisp_Object lispstream = Qnil; struct gcpro gcpro1; #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Vcurrent_compiled_function_annotation = Qnil; #endif GCPRO1 (lispstream); CHECK_STRING (string); get_string_range_byte (string, start, end, &startval, &endval, GB_HISTORICAL_STRING_BEHAVIOR); lispstream = make_lisp_string_input_stream (string, startval, endval - startval); read_objects = Qnil; tem = read0 (lispstream); /* Yeah, it's ugly. Gonna make something of it? At least our reader is reentrant ... */ tem = (Fcons (tem, make_int (bytecount_to_charcount (XSTRING_DATA (string), startval + Lstream_byte_count (XLSTREAM (lispstream)))))); Lstream_delete (XLSTREAM (lispstream)); UNGCPRO; return tem; } #ifdef LISP_BACKQUOTES static Lisp_Object backquote_unwind (Lisp_Object ptr) { /* used as unwind-protect function in read0() */ int *counter = (int *) get_opaque_ptr (ptr); if (--*counter < 0) *counter = 0; free_opaque_ptr (ptr); return Qnil; } #endif /* Use this for recursive reads, in contexts where internal tokens are not allowed. See also read1(). */ static Lisp_Object read0 (Lisp_Object readcharfun) { Lisp_Object val; val = read1 (readcharfun); if (CONSP (val) && UNBOUNDP (XCAR (val))) { Emchar c = XCHAR (XCDR (val)); free_cons (XCONS (val)); return Fsignal (Qinvalid_read_syntax, list1 (Fchar_to_string (make_char (c)))); } return val; } static Emchar read_escape (Lisp_Object readcharfun) { /* This function can GC */ Emchar c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); switch (c) { case 'a': return '\007'; case 'b': return '\b'; case 'd': return 0177; case 'e': return 033; case 'f': return '\f'; case 'n': return '\n'; case 'r': return '\r'; case 't': return '\t'; case 'v': return '\v'; case '\n': return -1; case 'M': c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); if (c != '-') error ("Invalid escape character syntax"); c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); if (c == '\\') c = read_escape (readcharfun); return c | 0200; /* Originally, FSF_KEYS provided a degree of FSF Emacs compatibility by defining character "modifiers" alt, super, hyper and shift to infest the characters (i.e. integers). However, this doesn't cut it for XEmacs 20, which distinguishes characters from integers. Without Mule, ?\H-a simply returns ?a because every character is clipped into 0-255. Under Mule it is much worse -- ?\H-a with FSF_KEYS produces an illegal character, and moves us to crash-land. For these reasons, FSF_KEYS hack is useless and without hope of ever working under XEmacs 20. */ #undef FSF_KEYS #ifdef FSF_KEYS #define alt_modifier (0x040000) #define super_modifier (0x080000) #define hyper_modifier (0x100000) #define shift_modifier (0x200000) /* fsf uses a different modifiers for meta and control. Possibly byte_compiled code will still work fsfmacs, though... --Stig #define ctl_modifier (0x400000) #define meta_modifier (0x800000) */ #define FSF_LOSSAGE(mask) \ if (fail_on_bucky_bit_character_escapes || \ ((c = readchar (readcharfun)) != '-')) \ error ("Invalid escape character syntax"); \ c = readchar (readcharfun); \ if (c < 0) \ signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); \ if (c == '\\') \ c = read_escape (readcharfun); \ return c | mask case 'S': FSF_LOSSAGE (shift_modifier); case 'H': FSF_LOSSAGE (hyper_modifier); case 'A': FSF_LOSSAGE (alt_modifier); case 's': FSF_LOSSAGE (super_modifier); #undef alt_modifier #undef super_modifier #undef hyper_modifier #undef shift_modifier #undef FSF_LOSSAGE #endif /* FSF_KEYS */ case 'C': c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); if (c != '-') error ("Invalid escape character syntax"); case '^': c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); if (c == '\\') c = read_escape (readcharfun); /* FSFmacs junk for non-ASCII controls. Not used here. */ if (c == '?') return 0177; else return c & (0200 | 037); case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': /* An octal escape, as in ANSI C. */ { REGISTER Emchar i = c - '0'; REGISTER int count = 0; while (++count < 3) { if ((c = readchar (readcharfun)) >= '0' && c <= '7') i = (i << 3) + (c - '0'); else { unreadchar (readcharfun, c); break; } } return i; } case 'x': /* A hex escape, as in ANSI C. */ { REGISTER Emchar i = 0; while (1) { c = readchar (readcharfun); /* Remember, can't use isdigit(), isalpha() etc. on Emchars */ if (c >= '0' && c <= '9') i = (i << 4) + (c - '0'); else if (c >= 'a' && c <= 'f') i = (i << 4) + (c - 'a') + 10; else if (c >= 'A' && c <= 'F') i = (i << 4) + (c - 'A') + 10; else { unreadchar (readcharfun, c); break; } } return i; } #ifdef MULE /* #### need some way of reading an extended character with an escape sequence. */ #endif default: return c; } } /* read symbol-constituent stuff into `Vread_buffer_stream'. */ static Bytecount read_atom_0 (Lisp_Object readcharfun, Emchar firstchar, int *saw_a_backslash) { /* This function can GC */ Emchar c = ((firstchar) >= 0 ? firstchar : readchar (readcharfun)); Lstream_rewind (XLSTREAM (Vread_buffer_stream)); *saw_a_backslash = 0; while (c > 040 /* #### - comma should be here as should backquote */ && !(c == '\"' || c == '\'' || c == ';' || c == '(' || c == ')' #ifndef LISP_FLOAT_TYPE /* If we have floating-point support, then we need to allow <digits><dot><digits>. */ || c =='.' #endif /* not LISP_FLOAT_TYPE */ || c == '[' || c == ']' || c == '#' )) { if (c == '\\') { c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); *saw_a_backslash = 1; } Lstream_put_emchar (XLSTREAM (Vread_buffer_stream), c); QUIT; c = readchar (readcharfun); } if (c >= 0) unreadchar (readcharfun, c); /* blasted terminating 0 */ Lstream_put_emchar (XLSTREAM (Vread_buffer_stream), 0); Lstream_flush (XLSTREAM (Vread_buffer_stream)); return Lstream_byte_count (XLSTREAM (Vread_buffer_stream)) - 1; } static Lisp_Object parse_integer (CONST Bufbyte *buf, Bytecount len, int base); static Lisp_Object read_atom (Lisp_Object readcharfun, Emchar firstchar, int uninterned_symbol) { /* This function can GC */ int saw_a_backslash; Bytecount len = read_atom_0 (readcharfun, firstchar, &saw_a_backslash); char *read_ptr = (char *) resizing_buffer_stream_ptr (XLSTREAM (Vread_buffer_stream)); /* Is it an integer? */ if (! (saw_a_backslash || uninterned_symbol)) { /* If a token had any backslashes in it, it is disqualified from being an integer or a float. This means that 123\456 is a symbol, as is \123 (which is the way (intern "123") prints). Also, if token was preceded by #:, it's always a symbol. */ char *p = read_ptr + len; char *p1 = read_ptr; if (*p1 == '+' || *p1 == '-') p1++; if (p1 != p) { int c; while (p1 != p && (c = *p1) >= '0' && c <= '9') p1++; #ifdef LISP_FLOAT_TYPE /* Integers can have trailing decimal points. */ if (p1 > read_ptr && p1 < p && *p1 == '.') p1++; #endif if (p1 == p) { /* It is an integer. */ #ifdef LISP_FLOAT_TYPE if (p1[-1] == '.') p1[-1] = '\0'; #endif #if 0 { int number = 0; if (sizeof (int) == sizeof (EMACS_INT)) number = atoi (read_buffer); else if (sizeof (long) == sizeof (EMACS_INT)) number = atol (read_buffer); else abort (); return make_int (number); } #else return parse_integer ((Bufbyte *) read_ptr, len, 10); #endif } } #ifdef LISP_FLOAT_TYPE if (isfloat_string (read_ptr)) return make_float (atof (read_ptr)); #endif } { Lisp_Object sym; if (uninterned_symbol) sym = (Fmake_symbol ((purify_flag) ? make_pure_pname ((Bufbyte *) read_ptr, len, 0) : make_string ((Bufbyte *) read_ptr, len))); else { /* intern will purecopy pname if necessary */ Lisp_Object name = make_string ((Bufbyte *) read_ptr, len); sym = Fintern (name, Qnil); if (SYMBOL_IS_KEYWORD (sym)) { /* the LISP way is to put keywords in their own package, but we don't have packages, so we do something simpler. Someday, maybe we'll have packages and then this will be reworked. --Stig. */ XSYMBOL (sym)->value = sym; } } return sym; } } static Lisp_Object parse_integer (CONST Bufbyte *buf, Bytecount len, int base) { CONST Bufbyte *lim = buf + len; CONST Bufbyte *p = buf; EMACS_UINT num = 0; int negativland = 0; if (*p == '-') { negativland = 1; p++; } else if (*p == '+') { p++; } if (p == lim) goto loser; for (; (p < lim) && (*p != '\0'); p++) { int c = *p; EMACS_UINT onum; if (isdigit (c)) c = c - '0'; else if (isupper (c)) c = c - 'A' + 10; else if (islower (c)) c = c - 'a' + 10; else goto loser; if (c < 0 || c >= base) goto loser; onum = num; num = num * base + c; if (num < onum) goto overflow; } { EMACS_INT int_result = negativland ? - (EMACS_INT) num : (EMACS_INT) num; Lisp_Object result = make_int (int_result); if (num && ((XINT (result) < 0) != negativland)) goto overflow; if (XINT (result) != int_result) goto overflow; return result; } overflow: return Fsignal (Qinvalid_read_syntax, list3 (build_translated_string ("Integer constant overflow in reader"), make_string (buf, len), make_int (base))); loser: return Fsignal (Qinvalid_read_syntax, list3 (build_translated_string ("Invalid integer constant in reader"), make_string (buf, len), make_int (base))); } static Lisp_Object read_integer (Lisp_Object readcharfun, int base) { /* This function can GC */ int saw_a_backslash; Bytecount len = read_atom_0 (readcharfun, -1, &saw_a_backslash); return (parse_integer (resizing_buffer_stream_ptr (XLSTREAM (Vread_buffer_stream)), ((saw_a_backslash) ? 0 /* make parse_integer signal error */ : len), base)); } static Lisp_Object read_bit_vector (Lisp_Object readcharfun) { unsigned_char_dynarr *dyn = Dynarr_new (unsigned_char); Emchar c; while (1) { c = readchar (readcharfun); if (c != '0' && c != '1') break; Dynarr_add (dyn, (unsigned char) (c - '0')); } if (c >= 0) unreadchar (readcharfun, c); return make_bit_vector_from_byte_vector (Dynarr_atp (dyn, 0), Dynarr_length (dyn)); } /* structures */ struct structure_type * define_structure_type (Lisp_Object type, int (*validate) (Lisp_Object data, Error_behavior errb), Lisp_Object (*instantiate) (Lisp_Object data)) { struct structure_type st; st.type = type; st.keywords = Dynarr_new (structure_keyword_entry); st.validate = validate; st.instantiate = instantiate; Dynarr_add (the_structure_type_dynarr, st); return Dynarr_atp (the_structure_type_dynarr, Dynarr_length (the_structure_type_dynarr) - 1); } void define_structure_type_keyword (struct structure_type *st, Lisp_Object keyword, int (*validate) (Lisp_Object keyword, Lisp_Object value, Error_behavior errb)) { struct structure_keyword_entry en; en.keyword = keyword; en.validate = validate; Dynarr_add (st->keywords, en); } static struct structure_type * recognized_structure_type (Lisp_Object type) { int i; for (i = 0; i < Dynarr_length (the_structure_type_dynarr); i++) { struct structure_type *st = Dynarr_atp (the_structure_type_dynarr, i); if (EQ (st->type, type)) return st; } return 0; } static Lisp_Object read_structure (Lisp_Object readcharfun) { Emchar c = readchar (readcharfun); Lisp_Object list = Qnil; Lisp_Object orig_list = Qnil; Lisp_Object already_seen = Qnil; int keyword_count; struct structure_type *st; struct gcpro gcpro1, gcpro2; GCPRO2 (orig_list, already_seen); if (c != '(') RETURN_UNGCPRO (continuable_syntax_error ("#s not followed by paren")); list = read_list (readcharfun, ')', 0, 0); orig_list = list; { int len = XINT (Flength (list)); if (len == 0) RETURN_UNGCPRO (continuable_syntax_error ("structure type not specified")); if (!(len & 1)) RETURN_UNGCPRO (continuable_syntax_error ("structures must have alternating keyword/value pairs")); } st = recognized_structure_type (XCAR (list)); if (!st) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("unrecognized structure type"), XCAR (list)))); list = Fcdr (list); keyword_count = Dynarr_length (st->keywords); while (!NILP (list)) { Lisp_Object keyword, value; int i; struct structure_keyword_entry *en = NULL; keyword = Fcar (list); list = Fcdr (list); value = Fcar (list); list = Fcdr (list); if (!NILP (memq_no_quit (keyword, already_seen))) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("structure keyword already seen"), keyword))); for (i = 0; i < keyword_count; i++) { en = Dynarr_atp (st->keywords, i); if (EQ (keyword, en->keyword)) break; } if (i == keyword_count) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("unrecognized structure keyword"), keyword))); if (en->validate && ! (en->validate) (keyword, value, ERROR_ME)) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list3 (build_translated_string ("invalid value for structure keyword"), keyword, value))); already_seen = Fcons (keyword, already_seen); } if (st->validate && ! (st->validate) (orig_list, ERROR_ME)) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("invalid structure initializer"), orig_list))); RETURN_UNGCPRO ((st->instantiate) (XCDR (orig_list))); } static Lisp_Object read_compiled_function (Lisp_Object readcharfun, int terminator); static Lisp_Object read_vector (Lisp_Object readcharfun, int terminator); /* Get the next character; filter out whitespace and comments */ static Emchar reader_nextchar (Lisp_Object readcharfun) { /* This function can GC */ Emchar c; retry: QUIT; c = readchar (readcharfun); if (c < 0) signal_error (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); switch (c) { default: { /* Ignore whitespace and control characters */ if (c <= 040) goto retry; return c; } case ';': { /* Comment */ while ((c = readchar (readcharfun)) >= 0 && c != '\n') QUIT; goto retry; } } } #if 0 static Lisp_Object list2_pure (int pure, Lisp_Object a, Lisp_Object b) { return pure ? pure_cons (a, pure_cons (b, Qnil)) : list2 (a, b); } #endif /* Read the next Lisp object from the stream READCHARFUN and return it. If the return value is a cons whose car is Qunbound, then read1() encountered a misplaced token (e.g. a right bracket, right paren, or dot followed by a non-number). To filter this stuff out, use read0(). */ static Lisp_Object read1 (Lisp_Object readcharfun) { Emchar c; retry: c = reader_nextchar (readcharfun); switch (c) { case '(': { #ifdef LISP_BACKQUOTES /* old backquote compatibility in lisp reader */ /* if this is disabled, then other code in eval.c must be enabled */ Emchar ch = reader_nextchar (readcharfun); switch (ch) { case '`': { Lisp_Object tem; int speccount = specpdl_depth (); ++old_backquote_flag; record_unwind_protect (backquote_unwind, make_opaque_ptr (&old_backquote_flag)); tem = read0 (readcharfun); unbind_to (speccount, Qnil); ch = reader_nextchar (readcharfun); if (ch != ')') { unreadchar (readcharfun, ch); return Fsignal (Qinvalid_read_syntax, list1 (build_string ("Weird old-backquote syntax"))); } return list2 (Qbacktick, tem); } case ',': { if (old_backquote_flag) { Lisp_Object tem, comma_type; ch = readchar (readcharfun); if (ch == '@') comma_type = Qcomma_at; else { if (ch >= 0) unreadchar (readcharfun, ch); comma_type = Qcomma; } tem = read0 (readcharfun); ch = reader_nextchar (readcharfun); if (ch != ')') { unreadchar (readcharfun, ch); return Fsignal (Qinvalid_read_syntax, list1 (build_string ("Weird old-backquote syntax"))); } return list2 (comma_type, tem); } else { unreadchar (readcharfun, ch); #if 0 return Fsignal (Qinvalid_read_syntax, list1 (build_string ("Comma outside of backquote"))); #else /* #### - yuck....but this is reverse compatible. */ /* mostly this is required by edebug, which does its own annotated reading. We need to have an annotated_read function that records (with markers) the buffer positions of the elements that make up lists, then that can be used in edebug and bytecomp and the check above can go back in. --Stig */ break; #endif } } default: unreadchar (readcharfun, ch); } /* switch(ch) */ #endif /* old backquote crap... */ return read_list (readcharfun, ')', 1, 1); } case '[': return read_vector (readcharfun, ']'); case ')': case ']': /* #### - huh? these don't do what they seem... */ return noseeum_cons (Qunbound, make_char (c)); case '.': { #ifdef LISP_FLOAT_TYPE /* If a period is followed by a number, then we should read it as a floating point number. Otherwise, it denotes a dotted pair. */ c = readchar (readcharfun); unreadchar (readcharfun, c); /* Can't use isdigit on Emchars */ if (c < '0' || c > '9') return noseeum_cons (Qunbound, make_char ('.')); /* Note that read_atom will loop at least once, assuring that we will not try to UNREAD two characters in a row. (I think this doesn't matter anymore because there should be no more danger in unreading multiple characters) */ return read_atom (readcharfun, '.', 0); #else /* ! LISP_FLOAT_TYPE */ return noseeum_cons (Qunbound, make_char ('.')); #endif /* ! LISP_FLOAT_TYPE */ } case '#': { c = readchar (readcharfun); switch (c) { #if 0 /* FSFmacs silly char-table syntax */ case '^': #endif #if 0 /* FSFmacs silly bool-vector syntax */ case '&': #endif /* "#["-- byte-code constant syntax */ /* purecons #[...] syntax */ case '[': return read_compiled_function (readcharfun, ']' /*, purify_flag */ ); /* "#:"-- gensym syntax */ case ':': return read_atom (readcharfun, -1, 1); /* #'x => (function x) */ case '\'': return list2 (Qfunction, read0 (readcharfun)); #if 0 /* RMS uses this syntax for fat-strings. If we use it for vectors, then obscure bugs happen. */ /* "#(" -- Scheme/CL vector syntax */ case '(': return read_vector (readcharfun, ')'); #endif #if 0 /* FSFmacs */ case '(': { Lisp_Object tmp; struct gcpro gcpro1; /* Read the string itself. */ tmp = read1 (readcharfun); if (!STRINGP (tmp)) { if (CONSP (tmp) && UNBOUNDP (XCAR (tmp))) free_cons (XCONS (tmp)); return Fsignal (Qinvalid_read_syntax, list1 (build_string ("#"))); } GCPRO1 (tmp); /* Read the intervals and their properties. */ while (1) { Lisp_Object beg, end, plist; Emchar ch; int invalid = 0; beg = read1 (readcharfun); if (CONSP (beg) && UNBOUNDP (XCAR (beg))) { ch = XCHAR (XCDR (beg)); free_cons (XCONS (beg)); if (ch == ')') break; else invalid = 1; } if (!invalid) { end = read1 (readcharfun); if (CONSP (end) && UNBOUNDP (XCAR (end))) { free_cons (XCONS (end)); invalid = 1; } } if (!invalid) { plist = read1 (readcharfun); if (CONSP (plist) && UNBOUNDP (XCAR (plist))) { free_cons (XCONS (plist)); invalid = 1; } } if (invalid) RETURN_UNGCPRO (Fsignal (Qinvalid_read_syntax, list2 (build_string ("invalid string property list"), XCDR (plist)))); Fset_text_properties (beg, end, plist, tmp); } UNGCPRO; return tmp; } #endif /* 0 */ case '@': { /* #@NUMBER is used to skip NUMBER following characters. That's used in .elc files to skip over doc strings and function definitions. */ int i, nskip = 0; /* Read a decimal integer. */ while ((c = readchar (readcharfun)) >= 0 && c >= '0' && c <= '9') nskip = (10 * nskip) + (c - '0'); if (c >= 0) unreadchar (readcharfun, c); /* FSF has code here that maybe caches the skipped string. See above for why this is totally losing. We handle this differently. */ /* Skip that many characters. */ for (i = 0; i < nskip && c >= 0; i++) c = readchar (readcharfun); goto retry; } case '$': return Vload_file_name_internal; /* bit vectors */ case '*': return read_bit_vector (readcharfun); /* #o10 => 8 -- octal constant syntax */ case 'o': return read_integer (readcharfun, 8); /* #xdead => 57005 -- hex constant syntax */ case 'x': return read_integer (readcharfun, 16); /* #b010 => 2 -- binary constant syntax */ case 'b': return read_integer (readcharfun, 2); /* #s(foobar key1 val1 key2 val2) -- structure syntax */ case 's': return read_structure (readcharfun); case '<': { unreadchar (readcharfun, c); return Fsignal (Qinvalid_read_syntax, list1 (build_string ("Cannot read unreadable object"))); } #ifdef FEATUREP_SYNTAX case '+': case '-': { Lisp_Object fexp, obj, tem; struct gcpro gcpro1, gcpro2; fexp = read0(readcharfun); obj = read0(readcharfun); /* the call to `featurep' may GC. */ GCPRO2(fexp, obj); tem = call1(Qfeaturep, fexp); UNGCPRO; if (c == '+' && NILP(tem)) goto retry; if (c == '-' && !NILP(tem)) goto retry; return obj; } #endif case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': /* Reader forms that can reuse previously read objects. */ { int n = 0; Lisp_Object found; /* 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. */ while (c >= '0' && c <= '9') { n *= 10; n += c - '0'; c = readchar (readcharfun); } found = assq_no_quit (make_int (n), read_objects); if (c == '=') { /* #n=object returns object, but associates it with n for #n#. */ Lisp_Object obj; if (CONSP (found)) return Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("Multiply defined symbol label"), make_int (n))); obj = read0 (readcharfun); read_objects = Fcons (Fcons (make_int (n), obj), read_objects); return obj; } else if (c == '#') { /* #n# returns a previously read object. */ if (CONSP (found)) return XCDR (found); else return Fsignal (Qinvalid_read_syntax, list2 (build_translated_string ("Undefined symbol label"), make_int (n))); } return Fsignal (Qinvalid_read_syntax, list1 (build_string ("#"))); } default: { unreadchar (readcharfun, c); return Fsignal (Qinvalid_read_syntax, list1 (build_string ("#"))); } } } /* Quote */ case '\'': return list2 (Qquote, read0 (readcharfun)); #ifdef LISP_BACKQUOTES case '`': { Lisp_Object tem; int speccount = specpdl_depth (); ++new_backquote_flag; record_unwind_protect (backquote_unwind, make_opaque_ptr (&new_backquote_flag)); tem = read0 (readcharfun); unbind_to (speccount, Qnil); return list2 (Qbackquote, tem); } case ',': { if (new_backquote_flag) { Lisp_Object comma_type = Qnil; int ch = readchar (readcharfun); if (ch == '@') comma_type = Qcomma_at; else if (ch == '.') comma_type = Qcomma_dot; else { if (ch >= 0) unreadchar (readcharfun, ch); comma_type = Qcomma; } return list2 (comma_type, read0 (readcharfun)); } else { /* YUCK. 99.999% backwards compatibility. The Right Thing(tm) is to signal an error here, because it's really invalid read syntax. Instead, this permits commas to begin symbols (unless they're inside backquotes). If an error is signalled here in the future, then commas should be invalid read syntax outside of backquotes anywhere they're found (i.e. they must be quoted in symbols) -- Stig */ return read_atom (readcharfun, c, 0); } } #endif case '?': { /* Evil GNU Emacs "character" (ie integer) syntax */ c = readchar (readcharfun); if (c < 0) return Fsignal (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); if (c == '\\') c = read_escape (readcharfun); return make_char (c); } case '\"': { /* String */ #ifdef I18N3 /* #### If the input stream is translating, then the string should be marked as translatable by setting its `string-translatable' property to t. .el and .elc files normally are translating input streams. See Fgettext() and print_internal(). */ #endif int cancel = 0; Lstream_rewind (XLSTREAM (Vread_buffer_stream)); while ((c = readchar (readcharfun)) >= 0 && c != '\"') { if (c == '\\') c = read_escape (readcharfun); /* c is -1 if \ newline has just been seen */ if (c == -1) { if (Lstream_byte_count (XLSTREAM (Vread_buffer_stream)) == 0) cancel = 1; } else Lstream_put_emchar (XLSTREAM (Vread_buffer_stream), c); QUIT; } if (c < 0) return Fsignal (Qend_of_file, list1 (READCHARFUN_MAYBE (readcharfun))); /* If purifying, and string starts with \ newline, return zero instead. This is for doc strings that we are really going to find in lib-src/DOC.nn.nn */ if (purify_flag && NILP (Vinternal_doc_file_name) && cancel) return Qzero; Lstream_flush (XLSTREAM (Vread_buffer_stream)); #if 0 /* FSFmacs defun hack */ if (read_pure) return make_pure_string (resizing_buffer_stream_ptr (XLSTREAM (Vread_buffer_stream)), Lstream_byte_count (XLSTREAM (Vread_buffer_stream))); else #endif return make_string (resizing_buffer_stream_ptr (XLSTREAM (Vread_buffer_stream)), Lstream_byte_count (XLSTREAM (Vread_buffer_stream))); } default: { /* Ignore whitespace and control characters */ if (c <= 040) goto retry; return read_atom (readcharfun, c, 0); } } } #ifdef LISP_FLOAT_TYPE #define LEAD_INT 1 #define DOT_CHAR 2 #define TRAIL_INT 4 #define E_CHAR 8 #define EXP_INT 16 int isfloat_string (CONST char *cp) { int state = 0; CONST Bufbyte *ucp = (CONST Bufbyte *) cp; if (*ucp == '+' || *ucp == '-') ucp++; if (*ucp >= '0' && *ucp <= '9') { state |= LEAD_INT; while (*ucp >= '0' && *ucp <= '9') ucp++; } if (*ucp == '.') { state |= DOT_CHAR; ucp++; } if (*ucp >= '0' && *ucp <= '9') { state |= TRAIL_INT; while (*ucp >= '0' && *ucp <= '9') ucp++; } if (*ucp == 'e' || *ucp == 'E') { state |= E_CHAR; ucp++; if ((*ucp == '+') || (*ucp == '-')) ucp++; } if (*ucp >= '0' && *ucp <= '9') { state |= EXP_INT; while (*ucp >= '0' && *ucp <= '9') ucp++; } return (((*ucp == 0) || (*ucp == ' ') || (*ucp == '\t') || (*ucp == '\n') || (*ucp == '\r') || (*ucp == '\f')) && (state == (LEAD_INT|DOT_CHAR|TRAIL_INT) || state == (DOT_CHAR|TRAIL_INT) || state == (LEAD_INT|E_CHAR|EXP_INT) || state == (LEAD_INT|DOT_CHAR|TRAIL_INT|E_CHAR|EXP_INT) || state == (DOT_CHAR|TRAIL_INT|E_CHAR|EXP_INT))); } #endif /* LISP_FLOAT_TYPE */ static void * sequence_reader (Lisp_Object readcharfun, Emchar terminator, void *state, void * (*conser) (Lisp_Object readcharfun, void *state, Charcount len)) { Charcount len; for (len = 0; ; len++) { Emchar ch; QUIT; ch = reader_nextchar (readcharfun); if (ch == terminator) return state; else unreadchar (readcharfun, ch); #ifdef FEATUREP_SYNTAX if (ch == ']') syntax_error ("\"]\" in a list"); else if (ch == ')') syntax_error ("\")\" in a vector"); #endif state = ((conser) (readcharfun, state, len)); } } struct read_list_state { Lisp_Object head; Lisp_Object tail; int length; int allow_dotted_lists; Emchar terminator; }; static void * read_list_conser (Lisp_Object readcharfun, void *state, Charcount len) { struct read_list_state *s = (struct read_list_state *) state; Lisp_Object elt; elt = read1 (readcharfun); if (CONSP (elt) && UNBOUNDP (XCAR (elt))) { Lisp_Object tem = elt; Emchar ch; elt = XCDR (elt); free_cons (XCONS (tem)); tem = Qnil; ch = XCHAR (elt); #ifdef FEATUREP_SYNTAX if (ch == s->terminator) /* deal with #+, #- reader macros */ { unreadchar (readcharfun, s->terminator); goto done; } else if (ch == ']') syntax_error ("']' in a list"); else if (ch == ')') syntax_error ("')' in a vector"); else #endif if (ch != '.') signal_simple_error ("BUG! Internal reader error", elt); else if (!s->allow_dotted_lists) syntax_error ("\".\" in a vector"); else { if (!NILP (s->tail)) XCDR (s->tail) = read0 (readcharfun); else s->head = read0 (readcharfun); elt = read1 (readcharfun); if (CONSP (elt) && UNBOUNDP (XCAR (elt))) { ch = XCHAR (XCDR (elt)); free_cons (XCONS (elt)); if (ch == s->terminator) { unreadchar (readcharfun, s->terminator); goto done; } } syntax_error (". in wrong context"); } } #if 0 /* FSFmacs defun hack, or something ... */ if (NILP (tail) && defun_hack && EQ (elt, Qdefun) && !read_pure) { record_unwind_protect (unreadpure, Qzero); read_pure = 1; } #endif #ifdef COMPILED_FUNCTION_ANNOTATION_HACK if (s->length == 1 && s->allow_dotted_lists && EQ (XCAR (s->head), Qfset)) { if (CONSP (elt) && EQ (XCAR (elt), Qquote) && CONSP (XCDR (elt))) Vcurrent_compiled_function_annotation = XCAR (XCDR (elt)); else Vcurrent_compiled_function_annotation = elt; } #endif elt = Fcons (elt, Qnil); if (!NILP (s->tail)) XCDR (s->tail) = elt; else s->head = elt; s->tail = elt; done: s->length++; return s; } #if 0 /* FSFmacs defun hack */ /* -1 for allow_dotted_lists means allow_dotted_lists and check for starting with defun and make structure pure. */ #endif static Lisp_Object read_list (Lisp_Object readcharfun, Emchar terminator, int allow_dotted_lists, int check_for_doc_references) { struct read_list_state s; struct gcpro gcpro1, gcpro2; #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Lisp_Object old_compiled_function_annotation = Vcurrent_compiled_function_annotation; #endif s.head = Qnil; s.tail = Qnil; s.length = 0; s.allow_dotted_lists = allow_dotted_lists; s.terminator = terminator; GCPRO2 (s.head, s.tail); sequence_reader (readcharfun, terminator, &s, read_list_conser); #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Vcurrent_compiled_function_annotation = old_compiled_function_annotation; #endif if ((purify_flag || load_force_doc_strings) && check_for_doc_references) { /* check now for any doc string references and record them for later. */ Lisp_Object tail; /* We might be dealing with an imperfect list so don't use LIST_LOOP */ for (tail = s.head; CONSP (tail); tail = XCDR (tail)) { Lisp_Object holding_cons = Qnil; { Lisp_Object elem = XCAR (tail); /* elem might be (#$ . INT) ... */ if (CONSP (elem) && EQ (XCAR (elem), Vload_file_name_internal)) holding_cons = tail; /* or it might be (quote (#$ . INT)) i.e. (quote . ((#$ . INT) . nil)) in the case of `autoload' (autoload evaluates its arguments, while `defvar', `defun', etc. don't). */ if (CONSP (elem) && EQ (XCAR (elem), Qquote) && CONSP (XCDR (elem))) { elem = XCAR (XCDR (elem)); if (CONSP (elem) && EQ (XCAR (elem), Vload_file_name_internal)) holding_cons = XCDR (XCAR (tail)); } } if (CONSP (holding_cons)) { if (purify_flag) { if (NILP (Vinternal_doc_file_name)) /* We have not yet called Snarf-documentation, so assume this file is described in the DOC file and Snarf-documentation will fill in the right value later. For now, replace the whole list with 0. */ XCAR (holding_cons) = Qzero; else /* We have already called Snarf-documentation, so make a relative file name for this file, so it can be found properly in the installed Lisp directory. We don't use Fexpand_file_name because that would make the directory absolute now. */ XCAR (XCAR (holding_cons)) = concat2 (build_string ("../lisp/"), Ffile_name_nondirectory (Vload_file_name_internal)); } else /* Not pure. Just add to Vload_force_doc_string_list, and the string will be filled in properly in load_force_doc_string_unwind(). */ Vload_force_doc_string_list = /* We pass the cons that holds the (#$ . INT) so we can modify it in-place. */ Fcons (holding_cons, Vload_force_doc_string_list); } } } UNGCPRO; return s.head; } static Lisp_Object read_vector (Lisp_Object readcharfun, Emchar terminator) { Lisp_Object tem; Lisp_Object *p; int len; int i; struct read_list_state s; struct gcpro gcpro1, gcpro2; s.head = Qnil; s.tail = Qnil; s.length = 0; s.allow_dotted_lists = 0; GCPRO2 (s.head, s.tail); sequence_reader (readcharfun, terminator, &s, read_list_conser); UNGCPRO; tem = s.head; len = XINT (Flength (tem)); #if 0 /* FSFmacs defun hack */ if (read_pure) s.head = make_pure_vector (len, Qnil); else #endif s.head = make_vector (len, Qnil); for (i = 0, p = &(XVECTOR_DATA (s.head)[0]); i < len; i++, p++) { struct Lisp_Cons *otem = XCONS (tem); #if 0 /* FSFmacs defun hack */ if (read_pure) tem = Fpurecopy (Fcar (tem)); else #endif tem = Fcar (tem); *p = tem; tem = otem->cdr; free_cons (otem); } return s.head; } static Lisp_Object read_compiled_function (Lisp_Object readcharfun, Emchar terminator) { /* Accept compiled functions at read-time so that we don't have to build them at load-time. */ Lisp_Object stuff; Lisp_Object make_byte_code_args[COMPILED_DOMAIN + 1]; struct gcpro gcpro1; int len; int iii; int saw_a_doc_ref = 0; /* Note: we tell read_list not to search for doc references because we need to handle the "doc reference" for the instructions and constants differently. */ stuff = read_list (readcharfun, terminator, 0, 0); len = XINT (Flength (stuff)); if (len < COMPILED_STACK_DEPTH + 1 || len > COMPILED_DOMAIN + 1) return continuable_syntax_error ("#[...] used with wrong number of elements"); for (iii = 0; CONSP (stuff); iii++) { struct Lisp_Cons *victim = XCONS (stuff); make_byte_code_args[iii] = Fcar (stuff); if ((purify_flag || load_force_doc_strings) && CONSP (make_byte_code_args[iii]) && EQ (XCAR (make_byte_code_args[iii]), Vload_file_name_internal)) { if (purify_flag && iii == COMPILED_DOC_STRING) { /* same as in read_list(). */ if (NILP (Vinternal_doc_file_name)) make_byte_code_args[iii] = Qzero; else XCAR (make_byte_code_args[iii]) = concat2 (build_string ("../lisp/"), Ffile_name_nondirectory (Vload_file_name_internal)); } else saw_a_doc_ref = 1; } stuff = Fcdr (stuff); free_cons (victim); } GCPRO1 (make_byte_code_args[0]); gcpro1.nvars = len; /* v18 or v19 bytecode file. Need to Ebolify. */ if (load_byte_code_version < 20 && VECTORP (make_byte_code_args[2])) ebolify_bytecode_constants (make_byte_code_args[2]); /* make-byte-code looks at purify_flag, which should have the same * value as our "read-pure" argument */ stuff = Fmake_byte_code (len, make_byte_code_args); XCOMPILED_FUNCTION (stuff)->flags.ebolified = (load_byte_code_version < 20); if (saw_a_doc_ref) Vload_force_doc_string_list = Fcons (stuff, Vload_force_doc_string_list); UNGCPRO; return stuff; } void init_lread (void) { Vvalues = Qnil; load_in_progress = 0; Vload_descriptor_list = Qnil; /* kludge: locate-file does not work for a null load-path, even if the file name is absolute. */ Vload_path = Fcons (build_string (""), Qnil); /* This used to get initialized in init_lread because all streams got closed when dumping occurs. This is no longer true -- Vread_buffer_stream is a resizing output stream, and there is no reason to close it at dump-time. Vread_buffer_stream is set to Qnil in vars_of_lread, and this will initialize it only once, at dump-time. */ if (NILP (Vread_buffer_stream)) Vread_buffer_stream = make_resizing_buffer_output_stream (); Vload_force_doc_string_list = Qnil; } void syms_of_lread (void) { DEFSUBR (Fread); DEFSUBR (Fread_from_string); DEFSUBR (Fload_internal); DEFSUBR (Flocate_file); DEFSUBR (Flocate_file_clear_hashing); DEFSUBR (Feval_buffer); DEFSUBR (Feval_region); defsymbol (&Qstandard_input, "standard-input"); defsymbol (&Qread_char, "read-char"); defsymbol (&Qcurrent_load_list, "current-load-list"); defsymbol (&Qload, "load"); defsymbol (&Qload_file_name, "load-file-name"); defsymbol (&Qlocate_file_hash_table, "locate-file-hash-table"); defsymbol (&Qfset, "fset"); #ifdef LISP_BACKQUOTES defsymbol (&Qbackquote, "backquote"); defsymbol (&Qbacktick, "`"); defsymbol (&Qcomma, ","); defsymbol (&Qcomma_at, ",@"); defsymbol (&Qcomma_dot, ",."); #endif } void structure_type_create (void) { the_structure_type_dynarr = Dynarr_new (structure_type); } void vars_of_lread (void) { DEFVAR_LISP ("values", &Vvalues /* List of values of all expressions which were read, evaluated and printed. Order is reverse chronological. */ ); DEFVAR_LISP ("standard-input", &Vstandard_input /* Stream for read to get input from. See documentation of `read' for possible values. */ ); Vstandard_input = Qt; DEFVAR_LISP ("load-path", &Vload_path /* *List of directories to search for files to load. Each element is a string (directory name) or nil (try default directory). Note that the elements of this list *may not* begin with "~", so you must call `expand-file-name' on them before adding them to this list. Initialized based on EMACSLOADPATH environment variable, if any, otherwise to default specified in by file `paths.h' when XEmacs was built. If there were no paths specified in `paths.h', then XEmacs chooses a default value for this variable by looking around in the file-system near the directory in which the XEmacs executable resides. */ ); Vload_path = Qnil; /* xxxDEFVAR_LISP ("dump-load-path", &Vdump_load_path, "*Location of lisp files to be used when dumping ONLY."); */ DEFVAR_BOOL ("load-in-progress", &load_in_progress /* Non-nil iff inside of `load'. */ ); DEFVAR_LISP ("after-load-alist", &Vafter_load_alist /* An alist of expressions to be evalled when particular files are loaded. Each element looks like (FILENAME FORMS...). When `load' is run and the file-name argument is FILENAME, the FORMS in the corresponding element are executed at the end of loading. FILENAME must match exactly! Normally FILENAME is the name of a library, with no directory specified, since that is how `load' is normally called. An error in FORMS does not undo the load, but does prevent execution of the rest of the FORMS. */ ); Vafter_load_alist = Qnil; DEFVAR_BOOL ("load-warn-when-source-newer", &load_warn_when_source_newer /* *Whether `load' should check whether the source is newer than the binary. If this variable is true, then when a `.elc' file is being loaded and the corresponding `.el' is newer, a warning message will be printed. */ ); load_warn_when_source_newer = 0; DEFVAR_BOOL ("load-warn-when-source-only", &load_warn_when_source_only /* *Whether `load' should warn when loading a `.el' file instead of an `.elc'. If this variable is true, then when `load' is called with a filename without an extension, and the `.elc' version doesn't exist but the `.el' version does, then a message will be printed. If an explicit extension is passed to `load', no warning will be printed. */ ); load_warn_when_source_only = 0; DEFVAR_BOOL ("load-ignore-elc-files", &load_ignore_elc_files /* *Whether `load' should ignore `.elc' files when a suffix is not given. This is normally used only to bootstrap the `.elc' files when building XEmacs. */ ); load_ignore_elc_files = 0; #ifdef LOADHIST DEFVAR_LISP ("load-history", &Vload_history /* Alist mapping source file names to symbols and features. Each alist element is a list that starts with a file name, except for one element (optional) that starts with nil and describes definitions evaluated from buffers not visiting files. The remaining elements of each list are symbols defined as functions or variables, and cons cells `(provide . FEATURE)' and `(require . FEATURE)'. */ ); Vload_history = Qnil; DEFVAR_LISP ("current-load-list", &Vcurrent_load_list /* Used for internal purposes by `load'. */ ); Vcurrent_load_list = Qnil; #endif DEFVAR_LISP ("load-file-name", &Vload_file_name /* Full name of file being loaded by `load'. */ ); Vload_file_name = Qnil; DEFVAR_LISP ("load-read-function", &Vload_read_function /* Function used by `load' and `eval-region' for reading expressions. The default is nil, which means use the function `read'. */ ); Vload_read_function = Qnil; DEFVAR_BOOL ("load-force-doc-strings", &load_force_doc_strings /* Non-nil means `load' should force-load all dynamic doc strings. This is useful when the file being loaded is a temporary copy. */ ); load_force_doc_strings = 0; /* See read_escape(). */ #if 0 /* Used to be named `puke-on-fsf-keys' */ DEFVAR_BOOL ("fail-on-bucky-bit-character-escapes", &fail_on_bucky_bit_character_escapes /* Whether `read' should signal an error when it encounters unsupported character escape syntaxes or just read them incorrectly. */ ); fail_on_bucky_bit_character_escapes = 0; #endif /* This must be initialized in init_lread otherwise it may start out with values saved when the image is dumped. */ staticpro (&Vload_descriptor_list); Vread_buffer_stream = Qnil; staticpro (&Vread_buffer_stream); /* Initialized in init_lread. */ staticpro (&Vload_force_doc_string_list); Vload_file_name_internal = Qnil; staticpro (&Vload_file_name_internal); Vload_file_name_internal_the_purecopy = Qnil; staticpro (&Vload_file_name_internal_the_purecopy); #ifdef COMPILED_FUNCTION_ANNOTATION_HACK Vcurrent_compiled_function_annotation = Qnil; staticpro (&Vcurrent_compiled_function_annotation); #endif /* So that early-early stuff will work */ Ffset (Qload, intern ("load-internal")); #ifdef FEATUREP_SYNTAX defsymbol (&Qfeaturep, "featurep"); Fprovide(intern("xemacs")); #ifdef INFODOCK Fprovide(intern("infodock")); #endif /* INFODOCK */ #endif /* FEATUREP_SYNTAX */ #ifdef LISP_BACKQUOTES old_backquote_flag = new_backquote_flag = 0; #endif #ifdef I18N3 Vfile_domain = Qnil; #endif read_objects = Qnil; staticpro (&read_objects); }