Adwaith-Rajesh/gperf
1
0
Fork 0
mirror of https://git.savannah.gnu.org/git/gperf.git synced 2025-12-02 21:19:24 +00:00
Code Activity

Move the input routines to class Input.

Browse Source
This commit is contained in:
Bruno Haible
2002-11-15 13:01:01 +00:00
parent 276e2fd949
commit ef7645289d
7 changed files with 504 additions and 401 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
ChangeLog
View File

@@ -1,3 +1,21 @@
2002-10-15 Bruno Haible <bruno@clisp.org>
* src/input.h: New file.
* src/input.cc: New file, extracted from key-list.cc.
* src/key-list.h (Key_List): Don't inherit from Read_Line.
(Key_List::get_special_input,
Key_List::save_include_src, Key_List::get_array_type,
Key_List::strcspn, Key_List::set_output_types): Remove methods.
* src/key-list.cc (Key_List::get_special_input,
Key_List::save_include_src, Key_List::get_array_type,
Key_List::strcspn, Key_List::set_output_types, parse_line): Move to
src/input.cc.
(Key_List::read_keys): Use Input::read_keys.
(Key_List::Key_List): Update.
* src/gen-perf.cc: Update.
* src/Makefile.in (OBJECTS): Add input.o.
(input.o): New rule.
2002-10-14 Bruno Haible <bruno@clisp.org>
* src/options.cc: Don't include "vector.h".

4
src/Makefile.in
View File

@@ -61,7 +61,7 @@ VPATH = $(srcdir)
OBJECTS = options.o main.o gen-perf.o key-list.o \
hash-table.o bool-array.o read-line.o vectors.o version.o \
keyword.o keyword-list.o output.o
keyword.o keyword-list.o output.o input.o
LIBS = ../lib/libgp.a @GPERF_LIBM@
CPPFLAGS = -I. -I$(srcdir)/../lib
@@ -116,6 +116,8 @@ keyword-list.o : keyword-list.cc keyword-list.h
$(CXX) $(CXXFLAGS) $(CPPFLAGS) -c $(srcdir)/keyword-list.cc
output.o : output.cc output.h
$(CXX) $(CXXFLAGS) $(CPPFLAGS) -c $(srcdir)/output.cc
input.o : input.cc input.h
$(CXX) $(CXXFLAGS) $(CPPFLAGS) -c $(srcdir)/input.cc
install : all force
$(MKINSTALLDIRS) $(DESTDIR)$(bindir)

2
src/gen-perf.cc
View File

@@ -41,7 +41,7 @@ Gen_Perf::Gen_Perf ()
int non_linked_length;
Vectors::ALPHA_SIZE = (option[SEVENBIT] ? 128 : 256);
Key_List::read_keys ();
read_keys ();
if (option[ORDER])
reorder ();
_num_done = 1;

347
src/input.cc Normal file
View File

@@ -0,0 +1,347 @@
/* Input routines.
Copyright (C) 1989-1998, 2002 Free Software Foundation, Inc.
Written by Douglas C. Schmidt <schmidt@ics.uci.edu>
and Bruno Haible <bruno@clisp.org>.
This file is part of GNU GPERF.
GNU GPERF 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 1, or (at your option)
any later version.
GNU GPERF 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 GNU GPERF; see the file COPYING. If not, write to the Free
Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111, USA. */
#include <stdio.h>
#include <stdlib.h> /* declares exit() */
#include <string.h> /* declares strncpy(), strchr() */
#include <limits.h> /* defines UCHAR_MAX etc. */
#include "options.h"
#include "input.h"
/* Gathers the input stream into a buffer until one of two things occur:
1. We read a '%' followed by a '%'
2. We read a '%' followed by a '}'
The first symbolizes the beginning of the keyword list proper,
The second symbolizes the end of the C source code to be generated
verbatim in the output file.
I assume that the keys are separated from the optional preceding struct
declaration by a consecutive % followed by either % or } starting in
the first column. The code below uses an expandible buffer to scan off
and return a pointer to all the code (if any) appearing before the delimiter. */
const char *
Input::get_special_input (char delimiter)
{
int size = 80;
char *buf = new char[size];
int c, i;
for (i = 0; (c = getchar ()) != EOF; i++)
{
if (c == '%')
{
if ((c = getchar ()) == delimiter)
{
while ((c = getchar ()) != '\n')
; /* discard newline */
if (i == 0)
return "";
else
{
buf[delimiter == '%' && buf[i - 2] == ';' ? i - 2 : i - 1] = '\0';
return buf;
}
}
else
buf[i++] = '%';
}
else if (i >= size) /* Yikes, time to grow the buffer! */
{
char *temp = new char[size *= 2];
int j;
for (j = 0; j < i; j++)
temp[j] = buf[j];
buf = temp;
}
buf[i] = c;
}
return 0; /* Problem here. */
}
/* Stores any C text that must be included verbatim into the
generated code output. */
const char *
Input::save_include_src ()
{
int c;
if ((c = getchar ()) != '%')
ungetc (c, stdin);
else if ((c = getchar ()) != '{')
{
fprintf (stderr, "internal error, %c != '{' on line %d in file %s", c, __LINE__, __FILE__);
exit (1);
}
else
return get_special_input ('}');
return "";
}
/* Determines from the input file whether the user wants to build a table
from a user-defined struct, or whether the user is content to simply
use the default array of keys. */
const char *
Input::get_array_type ()
{
return get_special_input ('%');
}
/* strcspn - find length of initial segment of S consisting entirely
of characters not from REJECT (borrowed from Henry Spencer's
ANSI string package, when GNU libc comes out I'll replace this...). */
#ifndef strcspn
inline int
Input::strcspn (const char *s, const char *reject)
{
const char *scan;
const char *rej_scan;
int count = 0;
for (scan = s; *scan; scan++)
{
for (rej_scan = reject; *rej_scan; rej_scan++)
if (*scan == *rej_scan)
return count;
count++;
}
return count;
}
#endif
/* Sets up the Return_Type, the Struct_Tag type and the Array_Type
based upon various user Options. */
void
Input::set_output_types ()
{
if (option[TYPE])
{
_array_type = get_array_type ();
if (!_array_type)
/* Something's wrong, but we'll catch it later on, in read_keys()... */
return;
/* Yow, we've got a user-defined type... */
int i = strcspn (_array_type, "{\n\0");
/* Remove trailing whitespace. */
while (i > 0 && strchr (" \t", _array_type[i-1]))
i--;
int struct_tag_length = i;
/* Set `struct_tag' to a naked "struct something". */
char *structtag = new char[struct_tag_length + 1];
strncpy (structtag, _array_type, struct_tag_length);
structtag[struct_tag_length] = '\0';
_struct_tag = structtag;
/* The return type of the lookup function is "struct something *".
No "const" here, because if !option[CONST], some user code might want
to modify the structure. */
char *rettype = new char[struct_tag_length + 3];
strncpy (rettype, _array_type, struct_tag_length);
rettype[struct_tag_length] = ' ';
rettype[struct_tag_length + 1] = '*';
rettype[struct_tag_length + 2] = '\0';
_return_type = rettype;
}
}
/* Extracts a key from an input line and creates a new KeywordExt_List for
it. */
static KeywordExt_List *
parse_line (const char *line, const char *delimiters)
{
if (*line == '"')
{
/* Parse a string in ANSI C syntax. */
char *key = new char[strlen(line)];
char *kp = key;
const char *lp = line + 1;
for (; *lp;)
{
char c = *lp;
if (c == '\0')
{
fprintf (stderr, "unterminated string: %s\n", line);
exit (1);
}
else if (c == '\\')
{
c = *++lp;
switch (c)
{
case '0': case '1': case '2': case '3':
case '4': case '5': case '6': case '7':
{
int code = 0;
int count = 0;
while (count < 3 && *lp >= '0' && *lp <= '7')
{
code = (code << 3) + (*lp - '0');
lp++;
count++;
}
if (code > UCHAR_MAX)
fprintf (stderr, "octal escape out of range: %s\n", line);
*kp = (char) code;
break;
}
case 'x':
{
int code = 0;
int count = 0;
lp++;
while ((*lp >= '0' && *lp <= '9')
|| (*lp >= 'A' && *lp <= 'F')
|| (*lp >= 'a' && *lp <= 'f'))
{
code = (code << 4)
+ (*lp >= 'A' && *lp <= 'F' ? *lp - 'A' + 10 :
*lp >= 'a' && *lp <= 'f' ? *lp - 'a' + 10 :
*lp - '0');
lp++;
count++;
}
if (count == 0)
fprintf (stderr, "hexadecimal escape without any hex digits: %s\n", line);
if (code > UCHAR_MAX)
fprintf (stderr, "hexadecimal escape out of range: %s\n", line);
*kp = (char) code;
break;
}
case '\\': case '\'': case '"':
*kp = c;
lp++;
break;
case 'n':
*kp = '\n';
lp++;
break;
case 't':
*kp = '\t';
lp++;
break;
case 'r':
*kp = '\r';
lp++;
break;
case 'f':
*kp = '\f';
lp++;
break;
case 'b':
*kp = '\b';
lp++;
break;
case 'a':
*kp = '\a';
lp++;
break;
case 'v':
*kp = '\v';
lp++;
break;
default:
fprintf (stderr, "invalid escape sequence in string: %s\n", line);
exit (1);
}
}
else if (c == '"')
break;
else
{
*kp = c;
lp++;
}
kp++;
}
lp++;
if (*lp != '\0')
{
if (strchr (delimiters, *lp) == NULL)
{
fprintf (stderr, "string not followed by delimiter: %s\n", line);
exit (1);
}
lp++;
}
return new KeywordExt_List (key, kp - key, option[TYPE] ? lp : "");
}
else
{
/* Not a string. Look for the delimiter. */
int len = strcspn (line, delimiters);
const char *rest;
if (line[len] == '\0')
rest = "";
else
/* Skip the first delimiter. */
rest = &line[len + 1];
return new KeywordExt_List (line, len, option[TYPE] ? rest : "");
}
}
void
Input::read_keys ()
{
char *ptr;
_include_src = save_include_src ();
set_output_types ();
/* Oops, problem with the input file. */
if (! (ptr = Read_Line::read_next_line ()))
{
fprintf (stderr, "No words in input file, did you forget to prepend %s or use -t accidentally?\n", "%%");
exit (1);
}
/* Read in all the keywords from the input file. */
const char *delimiter = option.get_delimiter ();
_head = parse_line (ptr, delimiter);
for (KeywordExt_List *temp = _head;
(ptr = Read_Line::read_next_line ()) && strcmp (ptr, "%%");
temp = temp->rest())
temp->rest() = parse_line (ptr, delimiter);
/* See if any additional C code is included at end of this file. */
if (ptr)
_additional_code = 1;
}

52
src/input.h Normal file
View File

@@ -0,0 +1,52 @@
/* This may look like C code, but it is really -*- C++ -*- */
/* Input routines.
Copyright (C) 1989-1998, 2002 Free Software Foundation, Inc.
Written by Douglas C. Schmidt <schmidt@ics.uci.edu>
and Bruno Haible <bruno@clisp.org>.
This file is part of GNU GPERF.
GNU GPERF 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 1, or (at your option)
any later version.
GNU GPERF 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 GNU GPERF; see the file COPYING. If not, write to the Free
Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111, USA. */
#ifndef input_h
#define input_h 1
#include "read-line.h"
#include "keyword-list.h"
class Input : private Read_Line
{
public:
void read_keys ();
private:
#ifndef strcspn
static int strcspn (const char *s, const char *reject);
#endif
void set_output_types ();
const char *get_array_type ();
const char *save_include_src ();
const char *get_special_input (char delimiter);
public:
const char *_array_type; /* Pointer to the type for word list. */
const char *_return_type; /* Pointer to return type for lookup function. */
const char *_struct_tag; /* Shorthand for user-defined struct tag type. */
const char *_include_src; /* C source code to be included verbatim. */
int _additional_code; /* True if any additional C code is included. */
KeywordExt_List *_head; /* Points to the head of the linked list. */
};
#endif

341
src/key-list.cc
View File

@@ -19,13 +19,12 @@ along with GNU GPERF; see the file COPYING. If not, write to the Free
Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111, USA. */
#include <stdio.h>
#include <string.h> /* declares strncpy(), strchr() */
#include <stdlib.h> /* declares malloc(), free(), abs(), exit(), abort() */
#include <limits.h> /* defines UCHAR_MAX etc. */
#include <stdlib.h> /* declares exit() */
#include <limits.h> /* defines INT_MIN, INT_MAX */
#include "options.h"
#include "read-line.h"
#include "hash-table.h"
#include "key-list.h"
#include "input.h"
#include "hash-table.h"
/* Make the hash table 8 times larger than the number of keyword entries. */
static const int TABLE_MULTIPLE = 10;
@@ -49,338 +48,31 @@ Key_List::~Key_List ()
}
}
/* Gathers the input stream into a buffer until one of two things occur:
1. We read a '%' followed by a '%'
2. We read a '%' followed by a '}'
The first symbolizes the beginning of the keyword list proper,
The second symbolizes the end of the C source code to be generated
verbatim in the output file.
I assume that the keys are separated from the optional preceding struct
declaration by a consecutive % followed by either % or } starting in
the first column. The code below uses an expandible buffer to scan off
and return a pointer to all the code (if any) appearing before the delimiter. */
const char *
Key_List::get_special_input (char delimiter)
{
int size = 80;
char *buf = new char[size];
int c, i;
for (i = 0; (c = getchar ()) != EOF; i++)
{
if (c == '%')
{
if ((c = getchar ()) == delimiter)
{
while ((c = getchar ()) != '\n')
; /* discard newline */
if (i == 0)
return "";
else
{
buf[delimiter == '%' && buf[i - 2] == ';' ? i - 2 : i - 1] = '\0';
return buf;
}
}
else
buf[i++] = '%';
}
else if (i >= size) /* Yikes, time to grow the buffer! */
{
char *temp = new char[size *= 2];
int j;
for (j = 0; j < i; j++)
temp[j] = buf[j];
buf = temp;
}
buf[i] = c;
}
return 0; /* Problem here. */
}
/* Stores any C text that must be included verbatim into the
generated code output. */
const char *
Key_List::save_include_src ()
{
int c;
if ((c = getchar ()) != '%')
ungetc (c, stdin);
else if ((c = getchar ()) != '{')
{
fprintf (stderr, "internal error, %c != '{' on line %d in file %s", c, __LINE__, __FILE__);
exit (1);
}
else
return get_special_input ('}');
return "";
}
/* Determines from the input file whether the user wants to build a table
from a user-defined struct, or whether the user is content to simply
use the default array of keys. */
const char *
Key_List::get_array_type ()
{
return get_special_input ('%');
}
/* strcspn - find length of initial segment of S consisting entirely
of characters not from REJECT (borrowed from Henry Spencer's
ANSI string package, when GNU libc comes out I'll replace this...). */
#ifndef strcspn
inline int
Key_List::strcspn (const char *s, const char *reject)
{
const char *scan;
const char *rej_scan;
int count = 0;
for (scan = s; *scan; scan++)
{
for (rej_scan = reject; *rej_scan; rej_scan++)
if (*scan == *rej_scan)
return count;
count++;
}
return count;
}
#endif
/* Sets up the Return_Type, the Struct_Tag type and the Array_Type
based upon various user Options. */
void
Key_List::set_output_types ()
{
if (option[TYPE])
{
_array_type = get_array_type ();
if (!_array_type)
/* Something's wrong, but we'll catch it later on, in read_keys()... */
return;
/* Yow, we've got a user-defined type... */
int i = strcspn (_array_type, "{\n\0");
/* Remove trailing whitespace. */
while (i > 0 && strchr (" \t", _array_type[i-1]))
i--;
int struct_tag_length = i;
/* Set `struct_tag' to a naked "struct something". */
char *structtag = new char[struct_tag_length + 1];
strncpy (structtag, _array_type, struct_tag_length);
structtag[struct_tag_length] = '\0';
_struct_tag = structtag;
/* The return type of the lookup function is "struct something *".
No "const" here, because if !option[CONST], some user code might want
to modify the structure. */
char *rettype = new char[struct_tag_length + 3];
strncpy (rettype, _array_type, struct_tag_length);
rettype[struct_tag_length] = ' ';
rettype[struct_tag_length + 1] = '*';
rettype[struct_tag_length + 2] = '\0';
_return_type = rettype;
}
}
/* Extracts a key from an input line and creates a new KeywordExt_List for
it. */
static KeywordExt_List *
parse_line (const char *line, const char *delimiters)
{
if (*line == '"')
{
/* Parse a string in ANSI C syntax. */
char *key = new char[strlen(line)];
char *kp = key;
const char *lp = line + 1;
for (; *lp;)
{
char c = *lp;
if (c == '\0')
{
fprintf (stderr, "unterminated string: %s\n", line);
exit (1);
}
else if (c == '\\')
{
c = *++lp;
switch (c)
{
case '0': case '1': case '2': case '3':
case '4': case '5': case '6': case '7':
{
int code = 0;
int count = 0;
while (count < 3 && *lp >= '0' && *lp <= '7')
{
code = (code << 3) + (*lp - '0');
lp++;
count++;
}
if (code > UCHAR_MAX)
fprintf (stderr, "octal escape out of range: %s\n", line);
*kp = (char) code;
break;
}
case 'x':
{
int code = 0;
int count = 0;
lp++;
while ((*lp >= '0' && *lp <= '9')
|| (*lp >= 'A' && *lp <= 'F')
|| (*lp >= 'a' && *lp <= 'f'))
{
code = (code << 4)
+ (*lp >= 'A' && *lp <= 'F' ? *lp - 'A' + 10 :
*lp >= 'a' && *lp <= 'f' ? *lp - 'a' + 10 :
*lp - '0');
lp++;
count++;
}
if (count == 0)
fprintf (stderr, "hexadecimal escape without any hex digits: %s\n", line);
if (code > UCHAR_MAX)
fprintf (stderr, "hexadecimal escape out of range: %s\n", line);
*kp = (char) code;
break;
}
case '\\': case '\'': case '"':
*kp = c;
lp++;
break;
case 'n':
*kp = '\n';
lp++;
break;
case 't':
*kp = '\t';
lp++;
break;
case 'r':
*kp = '\r';
lp++;
break;
case 'f':
*kp = '\f';
lp++;
break;
case 'b':
*kp = '\b';
lp++;
break;
case 'a':
*kp = '\a';
lp++;
break;
case 'v':
*kp = '\v';
lp++;
break;
default:
fprintf (stderr, "invalid escape sequence in string: %s\n", line);
exit (1);
}
}
else if (c == '"')
break;
else
{
*kp = c;
lp++;
}
kp++;
}
lp++;
if (*lp != '\0')
{
if (strchr (delimiters, *lp) == NULL)
{
fprintf (stderr, "string not followed by delimiter: %s\n", line);
exit (1);
}
lp++;
}
return new KeywordExt_List (key, kp - key, option[TYPE] ? lp : "");
}
else
{
/* Not a string. Look for the delimiter. */
int len = strcspn (line, delimiters);
const char *rest;
if (line[len] == '\0')
rest = "";
else
/* Skip the first delimiter. */
rest = &line[len + 1];
return new KeywordExt_List (line, len, option[TYPE] ? rest : "");
}
}
/* Reads in all keys from standard input and creates a linked list pointed
to by Head. This list is then quickly checked for ``links,'' i.e.,
to by _head. This list is then quickly checked for "links", i.e.,
unhashable elements possessing identical key sets and lengths. */
void
Key_List::read_keys ()
{
char *ptr;
Input inputter;
inputter.read_keys ();
_array_type = inputter._array_type;
_return_type = inputter._return_type;
_struct_tag = inputter._struct_tag;
_include_src = inputter._include_src;
_additional_code = inputter._additional_code;
_head = inputter._head;
_include_src = save_include_src ();
set_output_types ();
/* Oops, problem with the input file. */
if (! (ptr = Read_Line::read_next_line ()))
{
fprintf (stderr, "No words in input file, did you forget to prepend %s or use -t accidentally?\n", "%%");
exit (1);
}
/* Read in all the keywords from the input file. */
else
{
const char *delimiter = option.get_delimiter ();
KeywordExt_List *temp;
KeywordExt_List *trail = NULL;
_head = parse_line (ptr, delimiter);
_head->first()->init_selchars(this);
for (temp = _head;
(ptr = Read_Line::read_next_line ()) && strcmp (ptr, "%%");
temp = temp->rest())
for (temp = _head; temp; temp = temp->rest())
{
temp->rest() = parse_line (ptr, delimiter);
temp->rest()->first()->init_selchars(this);
temp->first()->init_selchars(this);
_total_keys++;
}
/* See if any additional C code is included at end of this file. */
if (ptr)
_additional_code = 1;
/* Hash table this number of times larger than keyword number. */
int table_size = (_list_len = _total_keys) * TABLE_MULTIPLE;
/* Table must be a power of 2 for the hash function scheme to work. */
@@ -450,7 +142,6 @@ Key_List::read_keys ()
exit (1);
}
}
}
/* Recursively merges two sorted lists together to form one sorted list. The
ordering criteria is by frequency of occurrence of elements in the key set
@@ -638,7 +329,7 @@ Key_List::dump ()
Key_List::Key_List ()
{
_total_keys = 1;
_total_keys = 0;
_max_key_len = INT_MIN;
_min_key_len = INT_MAX;
_array_type = 0;

9
src/key-list.h
View File

@@ -33,7 +33,7 @@ Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111, USA. */
#include "vectors.h"
#include "read-line.h"
class Key_List : private Read_Line, public Vectors
class Key_List : public Vectors
{
protected:
const char *_array_type; /* Pointer to the type for word list. */
@@ -55,16 +55,9 @@ protected:
private:
static int _determined[MAX_ALPHA_SIZE]; /* Used in function reorder, below. */
static int get_occurrence (KeywordExt *ptr);
#ifndef strcspn
static int strcspn (const char *s, const char *reject);
#endif
static int already_determined (KeywordExt *ptr);
static void set_determined (KeywordExt *ptr);
void set_output_types ();
void dump ();
const char *get_array_type ();
const char *save_include_src ();
const char *get_special_input (char delimiter);
KeywordExt_List *merge (KeywordExt_List *list1, KeywordExt_List *list2);
KeywordExt_List *merge_sort (KeywordExt_List *head);