/* This is the Assembler Pre-Processor
Copyright 1987, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
- 1999, 2000
+ 1999, 2000, 2001, 2002, 2003, 2006, 2007
Free Software Foundation, Inc.
This file is part of GAS, the GNU Assembler.
You should have received a copy of the GNU General Public License
along with GAS; see the file COPYING. If not, write to the Free
- Software Foundation, 59 Temple Place - Suite 330, Boston, MA
- 02111-1307, USA. */
+ Software Foundation, 51 Franklin Street - Fifth Floor, Boston, MA
+ 02110-1301, USA. */
-/* Modified by Allen Wirfs-Brock, Instantiations Inc 2/90 */
-/* App, the assembler pre-processor. This pre-processor strips out excess
- spaces, turns single-quoted characters into a decimal constant, and turns
- # <number> <filename> <garbage> into a .line <number>\n.file <filename>
- pair. This needs better error-handling. */
+/* Modified by Allen Wirfs-Brock, Instantiations Inc 2/90. */
+/* App, the assembler pre-processor. This pre-processor strips out
+ excess spaces, turns single-quoted characters into a decimal
+ constant, and turns the # in # <number> <filename> <garbage> into a
+ .linefile. This needs better error-handling. */
-#include <stdio.h>
-#include "as.h" /* For BAD_CASE() only */
+#include "as.h"
#if (__STDC__ != 1)
#ifndef const
flag_m68k_mri, because the two flags will be affected by the .mri
pseudo-op at different times. */
static int scrub_m68k_mri;
-#else
-#define scrub_m68k_mri 0
-#endif
/* The pseudo-op which switches in and out of MRI mode. See the
comment in do_scrub_chars. */
static const char mri_pseudo[] = ".mri 0";
+#else
+#define scrub_m68k_mri 0
+#endif
#if defined TC_ARM && defined OBJ_ELF
/* The pseudo-op for which we need to special-case `@' characters.
#define IS_LINE_COMMENT(c) (lex[c] == LEX_IS_LINE_COMMENT_START)
#define IS_NEWLINE(c) (lex[c] == LEX_IS_NEWLINE)
-static int process_escape PARAMS ((int));
+static int process_escape (int);
/* FIXME-soon: The entire lexer/parser thingy should be
built statically at compile time rather than dynamically
each and every time the assembler is run. xoxorich. */
void
-do_scrub_begin (m68k_mri)
- int m68k_mri ATTRIBUTE_UNUSED;
+do_scrub_begin (int m68k_mri ATTRIBUTE_UNUSED)
{
const char *p;
int c;
lex['"'] = LEX_IS_STRINGQUOTE;
#if ! defined (TC_HPPA) && ! defined (TC_I370)
- /* I370 uses single-quotes to delimit integer, float constants */
+ /* I370 uses single-quotes to delimit integer, float constants. */
lex['\''] = LEX_IS_ONECHAR_QUOTE;
#endif
/* Note that these override the previous defaults, e.g. if ';' is a
comment char, then it isn't a line separator. */
for (p = symbol_chars; *p; ++p)
- {
- lex[(unsigned char) *p] = LEX_IS_SYMBOL_COMPONENT;
- } /* declare symbol characters */
+ lex[(unsigned char) *p] = LEX_IS_SYMBOL_COMPONENT;
for (c = 128; c < 256; ++c)
lex[c] = LEX_IS_SYMBOL_COMPONENT;
#define tc_comment_chars comment_chars
#endif
for (p = tc_comment_chars; *p; p++)
- {
- lex[(unsigned char) *p] = LEX_IS_COMMENT_START;
- } /* declare comment chars */
+ lex[(unsigned char) *p] = LEX_IS_COMMENT_START;
for (p = line_comment_chars; *p; p++)
- {
- lex[(unsigned char) *p] = LEX_IS_LINE_COMMENT_START;
- } /* declare line comment chars */
+ lex[(unsigned char) *p] = LEX_IS_LINE_COMMENT_START;
for (p = line_separator_chars; *p; p++)
- {
- lex[(unsigned char) *p] = LEX_IS_LINE_SEPARATOR;
- } /* declare line separators */
+ lex[(unsigned char) *p] = LEX_IS_LINE_SEPARATOR;
#ifdef tc_parallel_separator_chars
/* This macro permits the processor to specify all characters which
separate parallel insns on the same line. */
for (p = tc_parallel_separator_chars; *p; p++)
- {
- lex[(unsigned char) *p] = LEX_IS_PARALLEL_SEPARATOR;
- } /* declare parallel separators */
+ lex[(unsigned char) *p] = LEX_IS_PARALLEL_SEPARATOR;
#endif
/* Only allow slash-star comments if slash is not in use.
FIXME: This isn't right. We should always permit them. */
if (lex['/'] == 0)
- {
- lex['/'] = LEX_IS_TWOCHAR_COMMENT_1ST;
- }
+ lex['/'] = LEX_IS_TWOCHAR_COMMENT_1ST;
#ifdef TC_M68K
if (m68k_mri)
lex[';'] = LEX_IS_COMMENT_START;
lex['*'] = LEX_IS_LINE_COMMENT_START;
/* The MRI documentation says '!' is LEX_IS_COMMENT_START, but
- then it can't be used in an expression. */
+ then it can't be used in an expression. */
lex['!'] = LEX_IS_LINE_COMMENT_START;
}
#endif
lex['|'] = LEX_IS_DOUBLEBAR_1ST;
#endif
#ifdef TC_D30V
- /* must do this is we want VLIW instruction with "->" or "<-" */
+ /* Must do this is we want VLIW instruction with "->" or "<-". */
lex['-'] = LEX_IS_SYMBOL_COMPONENT;
#endif
-} /* do_scrub_begin() */
+}
-/* Saved state of the scrubber */
+/* Saved state of the scrubber. */
static int state;
static int old_state;
static char *out_string;
state at the time .include is interpreted is completely unrelated.
That's why we have to save it all. */
-struct app_save {
+struct app_save
+{
int state;
int old_state;
char * out_string;
};
char *
-app_push ()
+app_push (void)
{
register struct app_save *saved;
}
void
-app_pop (arg)
- char *arg;
+app_pop (char *arg)
{
register struct app_save *saved = (struct app_save *) arg;
#endif
free (arg);
-} /* app_pop() */
+}
/* @@ This assumes that \n &c are the same on host and target. This is not
necessarily true. */
+
static int
-process_escape (ch)
- int ch;
+process_escape (int ch)
{
switch (ch)
{
This is the way the old code used to work. */
int
-do_scrub_chars (get, tostart, tolen)
- int (*get) PARAMS ((char *, int));
- char *tostart;
- int tolen;
+do_scrub_chars (int (*get) (char *, int), char *tostart, int tolen)
{
char *to = tostart;
char *toend = tostart + tolen;
char *fromend;
int fromlen;
register int ch, ch2 = 0;
+ /* Character that started the string we're working on. */
+ static char quotechar;
/*State 0: beginning of normal line
1: After first whitespace on line (flush more white)
2: After first non-white (opcode) on line (keep 1white)
3: after second white on line (into operands) (flush white)
- 4: after putting out a .line, put out digits
+ 4: after putting out a .linefile, put out digits
5: parsing a string, then go to old-state
6: putting out \ escape in a "d string.
- 7: After putting out a .appfile, put out string.
+ 7: no longer used
8: After putting out a .appfile string, flush until newline.
9: After seeing symbol char in state 3 (keep 1white after symchar)
10: After seeing whitespace in state 9 (keep white before symchar)
-1: output string in out_string and go to the state in old_state
-2: flush text until a '*' '/' is seen, then go to state old_state
#ifdef TC_V850
- 12: After seeing a dash, looking for a second dash as a start of comment.
+ 12: After seeing a dash, looking for a second dash as a start
+ of comment.
#endif
#ifdef DOUBLEBAR_PARALLEL
- 13: After seeing a vertical bar, looking for a second vertical bar as a parallel expression seperator.
+ 13: After seeing a vertical bar, looking for a second
+ vertical bar as a parallel expression separator.
+#endif
+#ifdef TC_IA64
+ 14: After seeing a `(' at state 0, looking for a `)' as
+ predicate.
+ 15: After seeing a `(' at state 1, looking for a `)' as
+ predicate.
+#endif
+#ifdef TC_Z80
+ 16: After seeing an 'a' or an 'A' at the start of a symbol
+ 17: After seeing an 'f' or an 'F' in state 16
#endif
*/
/* I added states 9 and 10 because the MIPS ECOFF assembler uses
constructs like ``.loc 1 20''. This was turning into ``.loc
120''. States 9 and 10 ensure that a space is never dropped in
- between characters which could appear in a identifier. Ian
+ between characters which could appear in an identifier. Ian
Taylor, ian@cygnus.com.
I added state 11 so that something like "Lfoo add %r25,%r26,%r27" works
I don't want to make such a significant change to the assembler's
memory usage. */
-#define PUT(pch) \
- do \
- { \
- *to++ = (pch); \
- if (to >= toend) \
- goto tofull; \
- } \
+#define PUT(pch) \
+ do \
+ { \
+ *to++ = (pch); \
+ if (to >= toend) \
+ goto tofull; \
+ } \
while (0)
if (saved_input != NULL)
while (1)
{
/* The cases in this switch end with continue, in order to
- branch back to the top of this while loop and generate the
- next output character in the appropriate state. */
+ branch back to the top of this while loop and generate the
+ next output character in the appropriate state. */
switch (state)
{
case -1:
ch = GET ();
if (ch == '"')
{
- UNGET (ch);
- if (scrub_m68k_mri)
- out_string = "\n\tappfile ";
- else
- out_string = "\n\t.appfile ";
- old_state = 7;
- state = -1;
- PUT (*out_string++);
+ PUT (' ');
+ PUT (ch);
+ quotechar = ch;
+ state = 5;
+ old_state = 8;
}
else
{
case 5:
/* We are going to copy everything up to a quote character,
- with special handling for a backslash. We try to
- optimize the copying in the simple case without using the
- GET and PUT macros. */
+ with special handling for a backslash. We try to
+ optimize the copying in the simple case without using the
+ GET and PUT macros. */
{
char *s;
int len;
for (s = from; s < fromend; s++)
{
ch = *s;
- /* This condition must be changed if the type of any
- other character can be LEX_IS_STRINGQUOTE. */
if (ch == '\\'
- || ch == '"'
- || ch == '\''
+ || ch == quotechar
|| ch == '\n')
break;
}
ch = GET ();
if (ch == EOF)
{
- as_warn (_("end of file in string; inserted '\"'"));
+ as_warn (_("end of file in string; '%c' inserted"), quotechar);
state = old_state;
UNGET ('\n');
- PUT ('"');
+ PUT (quotechar);
}
- else if (lex[ch] == LEX_IS_STRINGQUOTE)
+ else if (ch == quotechar)
{
state = old_state;
PUT (ch);
else if (scrub_m68k_mri && ch == '\n')
{
/* Just quietly terminate the string. This permits lines like
- bne label loop if we haven't reach end yet
- */
+ bne label loop if we haven't reach end yet. */
state = old_state;
UNGET (ch);
PUT ('\'');
PUT ('\\');
continue;
+ case EOF:
+ as_warn (_("end of file in string; '%c' inserted"), quotechar);
+ PUT (quotechar);
+ continue;
+
case '"':
case '\\':
case 'b':
case '6':
case '7':
break;
-#if defined(IGNORE_NONSTANDARD_ESCAPES) | defined(ONLY_STANDARD_ESCAPES)
+
default:
+#ifdef ONLY_STANDARD_ESCAPES
as_warn (_("unknown escape '\\%c' in string; ignored"), ch);
+#endif
break;
-#else /* ONLY_STANDARD_ESCAPES */
- default:
- /* Accept \x as x for any x */
- break;
-#endif /* ONLY_STANDARD_ESCAPES */
-
- case EOF:
- as_warn (_("end of file in string; '\"' inserted"));
- PUT ('"');
- continue;
}
PUT (ch);
continue;
- case 7:
- ch = GET ();
- state = 5;
- old_state = 8;
- if (ch == EOF)
- goto fromeof;
- PUT (ch);
- continue;
-
case 8:
do
- ch = GET ();
- while (ch != '\n' && ch != EOF);
- if (ch == EOF)
- goto fromeof;
+ if ((ch = GET ()) == EOF)
+ goto fromeof;
+ else
+ PUT (ch);
+ while (ch != '\n');
state = 0;
- PUT (ch);
continue;
+
+#ifdef DOUBLEBAR_PARALLEL
+ case 13:
+ ch = GET ();
+ if (ch != '|')
+ abort ();
+
+ /* Reset back to state 1 and pretend that we are parsing a
+ line from just after the first white space. */
+ state = 1;
+ PUT ('|');
+ continue;
+#endif
+#ifdef TC_Z80
+ case 16:
+ /* We have seen an 'a' at the start of a symbol, look for an 'f'. */
+ ch = GET ();
+ if (ch == 'f' || ch == 'F')
+ {
+ state = 17;
+ PUT (ch);
+ }
+ else
+ {
+ state = 9;
+ break;
+ }
+ case 17:
+ /* We have seen "af" at the start of a symbol,
+ a ' here is a part of that symbol. */
+ ch = GET ();
+ state = 9;
+ if (ch == '\'')
+ /* Change to avoid warning about unclosed string. */
+ PUT ('`');
+ else
+ UNGET (ch);
+ break;
+#endif
}
- /* OK, we are somewhere in states 0 through 4 or 9 through 11 */
+ /* OK, we are somewhere in states 0 through 4 or 9 through 11. */
/* flushchar: */
ch = GET ();
+#ifdef TC_IA64
+ if (ch == '(' && (state == 0 || state == 1))
+ {
+ state += 14;
+ PUT (ch);
+ continue;
+ }
+ else if (state == 14 || state == 15)
+ {
+ if (ch == ')')
+ {
+ state -= 14;
+ PUT (ch);
+ ch = GET ();
+ }
+ else
+ {
+ PUT (ch);
+ continue;
+ }
+ }
+#endif
+
recycle:
#if defined TC_ARM && defined OBJ_ELF
#ifdef TC_M68K
/* We want to have pseudo-ops which control whether we are in
- MRI mode or not. Unfortunately, since m68k MRI mode affects
- the scrubber, that means that we need a special purpose
- recognizer here. */
+ MRI mode or not. Unfortunately, since m68k MRI mode affects
+ the scrubber, that means that we need a special purpose
+ recognizer here. */
if (mri_state == NULL)
{
if ((state == 0 || state == 1)
else
{
/* We've read the entire pseudo-op. mips_last_ch is
- either '0' or '1' indicating whether to enter or
- leave MRI mode. */
+ either '0' or '1' indicating whether to enter or
+ leave MRI mode. */
do_scrub_begin (mri_last_ch == '1');
mri_state = NULL;
/* We continue handling the character as usual. The
- main gas reader must also handle the .mri pseudo-op
- to control expression parsing and the like. */
+ main gas reader must also handle the .mri pseudo-op
+ to control expression parsing and the like. */
}
}
#endif
if (lex[ch] == LEX_IS_COLON)
{
/* Only keep this white if there's no white *after* the
- colon. */
+ colon. */
ch2 = GET ();
UNGET (ch2);
if (!IS_WHITESPACE (ch2))
switch (state)
{
- case 0:
- state++;
- goto recycle; /* Punted leading sp */
case 1:
/* We can arrive here if we leave a leading whitespace
character at the beginning of a line. */
else
{
/* We know that ch is not ':', since we tested that
- case above. Therefore this is not a label, so it
- must be the opcode, and we've just seen the
- whitespace after it. */
+ case above. Therefore this is not a label, so it
+ must be the opcode, and we've just seen the
+ whitespace after it. */
state = 3;
}
UNGET (ch);
break;
/* This UNGET will ensure that we count newlines
- correctly. */
+ correctly. */
UNGET (ch2);
}
break;
case LEX_IS_STRINGQUOTE:
+ quotechar = ch;
if (state == 10)
{
- /* Preserve the whitespace in foo "bar" */
+ /* Preserve the whitespace in foo "bar". */
UNGET (ch);
state = 3;
PUT (' ');
/* PUT didn't jump out. We could just break, but we
- know what will happen, so optimize a bit. */
+ know what will happen, so optimize a bit. */
ch = GET ();
old_state = 3;
}
case LEX_IS_ONECHAR_QUOTE:
if (state == 10)
{
- /* Preserve the whitespace in foo 'b' */
+ /* Preserve the whitespace in foo 'b'. */
UNGET (ch);
state = 3;
PUT (' ');
ch = GET ();
}
while (ch != EOF && ch != '\n');
+
if (ch == EOF)
- {
- as_warn (_("end of file in comment; newline inserted"));
- }
+ as_warn (_("end of file in comment; newline inserted"));
+
state = 0;
PUT ('\n');
break;
#ifdef DOUBLEBAR_PARALLEL
case LEX_IS_DOUBLEBAR_1ST:
ch2 = GET ();
+ UNGET (ch2);
if (ch2 != '|')
- {
- UNGET (ch2);
- goto de_fault;
- }
- /* Reset back to state 1 and pretend that we are parsing a line from
- just after the first white space. */
- state = 1;
- PUT ('|');
+ goto de_fault;
+
+ /* Handle '||' in two states as invoking PUT twice might
+ result in the first one jumping out of this loop. We'd
+ then lose track of the state and one '|' char. */
+ state = 13;
PUT ('|');
break;
#endif
{
UNGET (ch2);
}
- } /* bad hack */
+ }
if (state == 0 || state == 1) /* Only comment at start of line. */
{
ch = GET ();
}
while (ch != EOF && IS_WHITESPACE (ch));
+
if (ch == EOF)
{
as_warn (_("end of file in comment; newline inserted"));
PUT ('\n');
break;
}
+
if (ch < '0' || ch > '9' || state != 0 || startch != '#')
{
/* Not a cpp line. */
old_state = 4;
state = -1;
if (scrub_m68k_mri)
- out_string = "\tappline ";
+ out_string = "\tlinefile ";
else
- out_string = "\t.appline ";
+ out_string = "\t.linefile ";
PUT (*out_string++);
break;
}
break;
}
+#ifdef TC_Z80
+ /* "af'" is a symbol containing '\''. */
+ if (state == 3 && (ch == 'a' || ch == 'A'))
+ {
+ state = 16;
+ PUT (ch);
+ ch = GET ();
+ if (ch == 'f' || ch == 'F')
+ {
+ state = 17;
+ PUT (ch);
+ break;
+ }
+ else
+ {
+ state = 9;
+ if (!IS_SYMBOL_COMPONENT (ch))
+ {
+ UNGET (ch);
+ break;
+ }
+ }
+ }
+#endif
if (state == 3)
state = 9;
/* This is a common case. Quickly copy CH and all the
- following symbol component or normal characters. */
+ following symbol component or normal characters. */
if (to + 1 < toend
&& mri_state == NULL
#if defined TC_ARM && defined OBJ_ELF
&& type != LEX_IS_SYMBOL_COMPONENT)
break;
}
+
if (s > from)
- {
- /* Handle the last character normally, for
- simplicity. */
- --s;
- }
+ /* Handle the last character normally, for
+ simplicity. */
+ --s;
+
len = s - from;
+
if (len > (toend - to) - 1)
len = (toend - to) - 1;
+
if (len > 0)
{
PUT (ch);
- if (len > 8)
- {
- memcpy (to, from, len);
- to += len;
- from += len;
- }
- else
- {
- switch (len)
- {
- case 8: *to++ = *from++;
- case 7: *to++ = *from++;
- case 6: *to++ = *from++;
- case 5: *to++ = *from++;
- case 4: *to++ = *from++;
- case 3: *to++ = *from++;
- case 2: *to++ = *from++;
- case 1: *to++ = *from++;
- }
- }
+ memcpy (to, from, len);
+ to += len;
+ from += len;
+ if (to >= toend)
+ goto tofull;
ch = GET ();
}
}
/* Some relatively `normal' character. */
if (state == 0)
{
- state = 11; /* Now seeing label definition */
+ state = 11; /* Now seeing label definition. */
}
else if (state == 1)
{
- state = 2; /* Ditto */
+ state = 2; /* Ditto. */
}
else if (state == 9)
{
- if (lex[ch] != LEX_IS_SYMBOL_COMPONENT)
+ if (!IS_SYMBOL_COMPONENT (ch))
state = 3;
}
else if (state == 10)
return to - tostart;
}
-/* end of app.c */
projects / deliverable / binutils-gdb.git / blobdiff