otherwise nothing is read. */
static bool
-lisp_file_lexically_bound_p (Lisp_Object readcharfun)
+lisp_file_lexically_bound_p (Lisp_Object readcharfun, bool *prefixes)
{
int ch = READCHAR;
+ /* We don't read package names as part of symbol_names by default. */
+ *prefixes = false;
+
if (ch == '#')
{
ch = READCHAR;
i--;
val[i] = '\0';
- if (strcmp (var, "lexical-binding") == 0)
- /* This is it... */
- {
- rv = (strcmp (val, "nil") != 0);
- break;
- }
+ /* PKG-FIXME Do this more elegantly? */
+ if (strcmp (var, "package-prefixes") == 0)
+ *prefixes = strcmp (val, "nil") == 0 ? false : true;
+ else if (strcmp (var, "lexical-binding") == 0)
+ rv = (strcmp (val, "nil") != 0);
}
}
}
else
{
- if (lisp_file_lexically_bound_p (Qget_file_char))
+ bool prefixes;
+ if (lisp_file_lexically_bound_p (Qget_file_char, &prefixes))
Fset (Qlexical_binding, Qt);
+ if (prefixes)
+ Fset (Qpackage_prefixes, Qt);
if (! version || version >= 22)
readevalloop (Qget_file_char, &input, hist_file_name,
specbind (Qstandard_output, tem);
record_unwind_protect_excursion ();
BUF_TEMP_SET_PT (XBUFFER (buf), BUF_BEGV (XBUFFER (buf)));
- specbind (Qlexical_binding, lisp_file_lexically_bound_p (buf) ? Qt : Qnil);
+ bool prefixes;
+ specbind (Qlexical_binding, lisp_file_lexically_bound_p (buf, &prefixes) ? Qt : Qnil);
+ specbind (Qpackage_prefixes, prefixes ? Qt : Qnil);
BUF_TEMP_SET_PT (XBUFFER (buf), BUF_BEGV (XBUFFER (buf)));
readevalloop (buf, 0, filename,
!NILP (printflag), unibyte, Qnil, Qnil, Qnil);
return make_pure_string (s, nchars, nbytes, multibyte);
}
+static bool
+is_symbol_constituent (int c)
+{
+ /* Symbols end at control characters like newlines or
+ tabs, or space of course. This if includes end of
+ input, where c < 0. */
+ if (c <= ' ')
+ return false;
+
+ /* Let symbols end at NO_BREAK_SPACE. */
+ if (c == NO_BREAK_SPACE)
+ return false;
+
+ /* Accept characters >= 128 as symbol constituents, like
+ unlauts and so on. */
+ if (c >= 128)
+ return true;
+
+ /* End reading when we reach a character that can not
+ be part of a symbol name, unless quoted. */
+ if (c == '"' || c == '\'' || c == ';' || c == '#'
+ || c == '(' || c == ')' || c == '[' || c == ']'
+ || c == '`' || c == ',')
+ return false;
+
+ return true;
+}
+
/* Read a Lisp object.
If LOCATE_SYMS is true, symbols are read with position. */
static Lisp_Object
EMACS_INT start_position = readchar_offset - 1;
/* PKG-FIXME: This is too complicated. */
+ /* PKG-FIXME: Check package-prefixes binding working. */
/* Remember where package prefixes end in COLON, which
will be set to the first colon we find. NCOLONS is the
char *colon = NULL;
int ncolons = 0;
- /* True means last character read was a backslash. */
+ /* True if last character read was a backslash. */
bool last_was_backslash = false;
- bool in_vertical_bar = false;
+ /* True if \ for escaping appeared. */
bool any_quoted = false;
for (;;)
{
- if (c == ':'
- // This is actually \: or |...:
- && !last_was_backslash && !in_vertical_bar)
+ eassert (is_symbol_constituent (c) || last_was_backslash);
+
+ /* Treat ':' as package prefix, unless someone says we
+ should't, or it is escaped by a preceding '\\' or
+ inside a multi-escape. Note that we don't land here
+ for #:. */
+ if (c == ':' && !last_was_backslash && !NILP (Vpackage_prefixes))
{
/* Remember where the first : is. */
if (colon == NULL)
colon = p;
++ncolons;
- if (!read_emacs_syntax)
- {
- /* #:xyz should not contain a colon unless in Emacs
- original syntax. */
- if (uninterned_symbol)
- invalid_syntax ("colon in uninterned symbol", readcharfun);
-
- /* Up to two colons are allowed if they are
- consecutive. PKG-FIXME check consecutive :. */
- if (ncolons > 2)
- invalid_syntax ("too many colons", readcharfun);
- }
+ /* #:xyz should not contain a colon unless in Emacs
+ original syntax. */
+ if (uninterned_symbol)
+ invalid_syntax ("colon in uninterned symbol", readcharfun);
+
+ /* Up to two colons are allowed if they are
+ consecutive. PKG-FIXME check consecutive :. */
+ if (ncolons > 2)
+ invalid_syntax ("too many colons", readcharfun);
}
- /* Handle unquote backslash and bar . */
- if (!last_was_backslash)
+ /* unescaped backslash. Remember that we have seen it. */
+ if (c == '\\' && !last_was_backslash)
{
- /* Unquoted backslash: The first backslash is not part
- of the symbol name. \\ gives a single \ in the
- symbol. */
- if (c == '\\')
- {
- c = READCHAR;
- if (c < 0)
- end_of_file_error ();
- last_was_backslash = true;
- any_quoted = true;
- continue;
- }
-
- /* Unquoted vertical bar. Begin or end multi-escape,
- unless in Emacs syntax. In either case, proceed
- with next char, the bar is not part of the
- name. */
- if (c == '|' && !read_emacs_syntax)
- {
- c = READCHAR;
- if (c < 0)
- end_of_file_error ();
- in_vertical_bar = !in_vertical_bar;
- continue;
- }
+ any_quoted = true;
+ last_was_backslash = true;
+ c = READCHAR;
+ if (c < 0)
+ invalid_syntax ("eof in single-escape", readcharfun);
+ continue;
}
last_was_backslash = false;
/* Proceed with the next character. */
c = READCHAR;
- /* If in |...| everything is quoted by default, except |.
- An unquoted bar ends the symbol, and is not part of the
- symbol. */
- if (in_vertical_bar)
- {
- eassert (!read_emacs_syntax);
- if (c < 0)
- end_of_file_error ();
- if (c == '|')
- {
- c = READCHAR;
- break;
- }
- }
- else
- {
- /* Symbols end at control characters like newlines or
- tabs, or space of course. This if includes end of
- input, where c < 0. */
- if (c <= ' ')
- break;
-
- /* Let symbols end at NO_BREAK_SPACE. */
- if (c == NO_BREAK_SPACE)
- break;
-
- /* Accept characters >= 128 as symbol constituents, like
- unlauts and so on. */
- if (c >= 128)
- continue;
-
- /* End reading when we reach a character that can not
- be part of a symbol name, unless quoted. */
- if (c == '"' || c == '\'' || c == ';' || c == '#'
- || c == '(' || c == ')' || c == '[' || c == ']'
- || c == '`' || c == ',')
- break;
- }
+ /* Symbols end at control characters like newlines or
+ tabs, or space of course. This if includes end of
+ input, where c < 0. */
+ if (c <= ' ')
+ break;
+
+ /* Let symbols end at NO_BREAK_SPACE. */
+ if (c == NO_BREAK_SPACE)
+ break;
+
+ /* Accept characters >= 128 as symbol constituents, like
+ unlauts and so on. */
+ if (c >= 128)
+ continue;
+
+ /* End reading when we reach a character that can not
+ be part of a symbol name, unless quoted. */
+ if (c == '"' || c == '\'' || c == ';' || c == '#'
+ || c == '(' || c == ')' || c == '[' || c == ']'
+ || c == '`' || c == ',')
+ break;
}
+ eassert (!is_symbol_constituent (c));
+ /* c maybe -1 here, hut we can unread EOF. */
*p = 0;
UNREAD (c);
/* The start of the symbol, If a package prefix is present,
- the start of the symbol-name part. */
- char *symbol_start;
+ set to the start of the symbol-name part later on. */
+ char *symbol_start = read_buffer;
+ const char *symbol_end = p;
/* Package for the package prefix, if there is one, or nil
if there is none. */
/* If a package prefix was found, determine the package it
names. It is an error if a package of that name does not
- exist, or ':' is used for an internal symbol. */
- if (colon)
+ exist, or ':' is used for an internal symbol.
+
+ If we don't want to recognize ':' as a package indicator,
+ nevertheless handle keywords. */
+ if (NILP (Vpackage_prefixes))
{
- /* Package name is in read_buffer, colon + ncolons is the
+ if (*symbol_start == ':')
+ {
+ ++symbol_start;
+ package = Ffind_package (Qkeyword);
+ eassert (!NILP (package));
+ }
+ }
+ else if (colon)
+ {
+ /* PACKAGE name is in read_buffer, colon + ncolons is the
start of the symbol name. */
*colon = 0;
/* Symbol name starts after the package prefix. */
symbol_start = colon + ncolons;
}
- else
- symbol_start = read_buffer;
- const ptrdiff_t symbol_nbytes = p - symbol_start;
- /* This could be a number after all. But not if empty, and not
- if in |...|, and not if any quoted characters were found,
- or a package prefix was found, or we have #:xyz. */
+ /* This could be a number after all. But not if empty, and
+ not if anything was quoted. or a package prefix was found,
+ or we have #:xyz. */
+ const ptrdiff_t symbol_nbytes = symbol_end - symbol_start;
if (!any_quoted
- && !in_vertical_bar
- && !colon
- && !uninterned_symbol)
+ && !uninterned_symbol
+ && NILP (package)
+ && symbol_end != symbol_start)
{
- char c0 = symbol_start[0];
+ char c0 = *symbol_start;
if (((c0 >= '0' && c0 <= '9') || c0 == '.' || c0 == '-' || c0 == '+')
&& !skip_shorthand)
{
ptrdiff_t len;
- Lisp_Object result = string_to_number (read_buffer, 10, &len);
+ /* 10 as base because the other bases require a #, and
+ don't land here. */
+ Lisp_Object result = string_to_number (symbol_start, 10, &len);
if (!NILP (result) && len == symbol_nbytes)
{
obj = result;
doc: /* Non-nil means not to load a .eln file when a .elc was requested. */);
load_no_native = false;
- DEFVAR_BOOL ("read-emacs-syntax", read_emacs_syntax,
- doc: /* Non-nil means don't treat ':' or '|' specially in symbols. */);
- read_emacs_syntax = true;
-
/* Vsource_directory was initialized in init_lread. */
DEFSYM (Qcurrent_load_list, "current-load-list");
projects / emacs.git / commitdiff