joachim99@52: /* Shared definitions for GNU DIFF
joachim99@52: 
joachim99@52:    Modified for KDiff3 by Joachim Eibl 2003.
joachim99@53:    The original file was part of GNU DIFF.
joachim99@52: 
joachim99@52:    Copyright (C) 1988, 1989, 1991, 1992, 1993, 1994, 1995, 1998, 2001,
joachim99@52:    2002 Free Software Foundation, Inc.
joachim99@52: 
joachim99@52:    GNU DIFF is free software; you can redistribute it and/or modify
joachim99@52:    it under the terms of the GNU General Public License as published by
joachim99@52:    the Free Software Foundation; either version 2, or (at your option)
joachim99@52:    any later version.
joachim99@52: 
joachim99@52:    GNU DIFF is distributed in the hope that it will be useful,
joachim99@52:    but WITHOUT ANY WARRANTY; without even the implied warranty of
joachim99@52:    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
joachim99@52:    GNU General Public License for more details.
joachim99@52: 
joachim99@52:    You should have received a copy of the GNU General Public License
joachim99@52:    along with this program; see the file COPYING.
joachim99@52:    If not, write to the Free Software Foundation,
joachim99@52:    59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
joachim99@52: 
joachim99@53: #ifndef GNUDIFF_DIFF_H
joachim99@53: #define GNUDIFF_DIFF_H
joachim99@53: 
joachim99@52: #include "gnudiff_system.h"
joachim99@52: 
joachim99@52: #include <stdio.h>
joachim99@52: 
joachim99@52: #define TAB_WIDTH 8
joachim99@52: 
joachim99@53: class GnuDiff
joachim99@52: {
joachim99@53: public:
joachim99@52: /* What kind of changes a hunk contains.  */
joachim99@52: enum changes
joachim99@52: {
joachim99@52:   /* No changes: lines common to both files.  */
joachim99@52:   UNCHANGED,
joachim99@52: 
joachim99@52:   /* Deletes only: lines taken from just the first file.  */
joachim99@52:   OLD,
joachim99@52: 
joachim99@52:   /* Inserts only: lines taken from just the second file.  */
joachim99@52:   NEW,
joachim99@52: 
joachim99@52:   /* Both deletes and inserts: a hunk containing both old and new lines.  */
joachim99@52:   CHANGED
joachim99@52: };
joachim99@52: 
joachim99@52: /* Variables for command line options */
joachim99@52: 
joachim99@52: enum output_style
joachim99@52: {
joachim99@52:   /* No output style specified.  */
joachim99@52:   OUTPUT_UNSPECIFIED,
joachim99@52: 
joachim99@52:   /* Default output style.  */
joachim99@52:   OUTPUT_NORMAL,
joachim99@52: 
joachim99@52:   /* Output the differences with lines of context before and after (-c).  */
joachim99@52:   OUTPUT_CONTEXT,
joachim99@52: 
joachim99@52:   /* Output the differences in a unified context diff format (-u).  */
joachim99@52:   OUTPUT_UNIFIED,
joachim99@52: 
joachim99@52:   /* Output the differences as commands suitable for `ed' (-e).  */
joachim99@52:   OUTPUT_ED,
joachim99@52: 
joachim99@52:   /* Output the diff as a forward ed script (-f).  */
joachim99@52:   OUTPUT_FORWARD_ED,
joachim99@52: 
joachim99@52:   /* Like -f, but output a count of changed lines in each "command" (-n).  */
joachim99@52:   OUTPUT_RCS,
joachim99@52: 
joachim99@52:   /* Output merged #ifdef'd file (-D).  */
joachim99@52:   OUTPUT_IFDEF,
joachim99@52: 
joachim99@52:   /* Output sdiff style (-y).  */
joachim99@52:   OUTPUT_SDIFF
joachim99@52: };
joachim99@52: 
joachim99@52: /* True for output styles that are robust,
joachim99@52:    i.e. can handle a file that ends in a non-newline.  */
joachim99@52: #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
joachim99@52: 
joachim99@53: enum output_style output_style;
joachim99@52: 
joachim99@52: /* Nonzero if output cannot be generated for identical files.  */
joachim99@53: bool no_diff_means_no_output;
joachim99@52: 
joachim99@52: /* Number of lines of context to show in each set of diffs.
joachim99@52:    This is zero when context is not to be shown.  */
joachim99@53: lin context;
joachim99@52: 
joachim99@52: /* Consider all files as text files (-a).
joachim99@52:    Don't interpret codes over 0177 as implying a "binary file".  */
joachim99@53: bool text;
joachim99@52: 
joachim99@52: /* Number of lines to keep in identical prefix and suffix.  */
joachim99@53: lin horizon_lines;
joachim99@52: 
joachim99@52: /* The significance of white space during comparisons.  */
joachim99@53: enum
joachim99@52: {
joachim99@52:   /* All white space is significant (the default).  */
joachim99@52:   IGNORE_NO_WHITE_SPACE,
joachim99@52: 
joachim99@52:   /* Ignore changes due to tab expansion (-E).  */
joachim99@52:   IGNORE_TAB_EXPANSION,
joachim99@52: 
joachim99@52:   /* Ignore changes in horizontal white space (-b).  */
joachim99@52:   IGNORE_SPACE_CHANGE,
joachim99@52: 
joachim99@52:   /* Ignore all horizontal white space (-w).  */
joachim99@52:   IGNORE_ALL_SPACE
joachim99@52: } ignore_white_space;
joachim99@52: 
joachim99@52: /* Ignore changes that affect only blank lines (-B).  */
joachim99@53: bool ignore_blank_lines;
joachim99@52: 
joachim99@52: /* Ignore changes that affect only numbers. (J. Eibl)  */
joachim99@53: bool bIgnoreNumbers;
joachim99@53: bool bIgnoreWhiteSpace;
joachim99@52: 
joachim99@52: /* Files can be compared byte-by-byte, as if they were binary.
joachim99@52:    This depends on various options.  */
joachim99@53: bool files_can_be_treated_as_binary;
joachim99@52: 
joachim99@52: /* Ignore differences in case of letters (-i).  */
joachim99@53: bool ignore_case;
joachim99@52: 
joachim99@52: /* Ignore differences in case of letters in file names.  */
joachim99@53: bool ignore_file_name_case;
joachim99@52: 
joachim99@52: /* File labels for `-c' output headers (--label).  */
joachim99@53: char *file_label[2];
joachim99@52: 
joachim99@52: /* Regexp to identify function-header lines (-F).  */
joachim99@53: //struct re_pattern_buffer function_regexp;
joachim99@52: 
joachim99@52: /* Ignore changes that affect only lines matching this regexp (-I).  */
joachim99@53: //struct re_pattern_buffer ignore_regexp;
joachim99@52: 
joachim99@52: /* Say only whether files differ, not how (-q).  */
joachim99@53: bool brief;
joachim99@52: 
joachim99@52: /* Expand tabs in the output so the text lines up properly
joachim99@52:    despite the characters added to the front of each line (-t).  */
joachim99@53: bool expand_tabs;
joachim99@52: 
joachim99@52: /* Use a tab in the output, rather than a space, before the text of an
joachim99@52:    input line, so as to keep the proper alignment in the input line
joachim99@52:    without changing the characters in it (-T).  */
joachim99@53: bool initial_tab;
joachim99@52: 
joachim99@52: /* Remove trailing carriage returns from input.  */
joachim99@53: bool strip_trailing_cr;
joachim99@52: 
joachim99@52: /* In directory comparison, specify file to start with (-S).
joachim99@52:    This is used for resuming an aborted comparison.
joachim99@52:    All file names less than this name are ignored.  */
joachim99@53: char const *starting_file;
joachim99@52: 
joachim99@52: /* Pipe each file's output through pr (-l).  */
joachim99@53: bool paginate;
joachim99@52: 
joachim99@52: /* Line group formats for unchanged, old, new, and changed groups.  */
joachim99@53: char const *group_format[CHANGED + 1];
joachim99@52: 
joachim99@52: /* Line formats for unchanged, old, and new lines.  */
joachim99@53: char const *line_format[NEW + 1];
joachim99@52: 
joachim99@52: /* If using OUTPUT_SDIFF print extra information to help the sdiff filter.  */
joachim99@53: bool sdiff_merge_assist;
joachim99@52: 
joachim99@52: /* Tell OUTPUT_SDIFF to show only the left version of common lines.  */
joachim99@53: bool left_column;
joachim99@52: 
joachim99@52: /* Tell OUTPUT_SDIFF to not show common lines.  */
joachim99@53: bool suppress_common_lines;
joachim99@52: 
joachim99@52: /* The half line width and column 2 offset for OUTPUT_SDIFF.  */
joachim99@53: unsigned int sdiff_half_width;
joachim99@53: unsigned int sdiff_column2_offset;
joachim99@52: 
joachim99@52: /* String containing all the command options diff received,
joachim99@52:    with spaces between and at the beginning but none at the end.
joachim99@52:    If there were no options given, this string is empty.  */
joachim99@53: char *switch_string;
joachim99@52: 
joachim99@52: /* Use heuristics for better speed with large files with a small
joachim99@52:    density of changes.  */
joachim99@53: bool speed_large_files;
joachim99@52: 
joachim99@52: /* Patterns that match file names to be excluded.  */
joachim99@53: struct exclude *excluded;
joachim99@52: 
joachim99@52: /* Don't discard lines.  This makes things slower (sometimes much
joachim99@52:    slower) but will find a guaranteed minimal set of changes.  */
joachim99@53: bool minimal;
joachim99@52: 
joachim99@52: /* Name of program the user invoked (for error messages).  */
joachim99@53: char *program_name;
joachim99@52: 
joachim99@52: /* The strftime format to use for time strings.  */
joachim99@53: char const *time_format;
joachim99@52: 
joachim99@52: /* The result of comparison is an "edit script": a chain of `struct change'.
joachim99@52:    Each `struct change' represents one place where some lines are deleted
joachim99@52:    and some are inserted.
joachim99@52: 
joachim99@52:    LINE0 and LINE1 are the first affected lines in the two files (origin 0).
joachim99@52:    DELETED is the number of lines deleted here from file 0.
joachim99@52:    INSERTED is the number of lines inserted here in file 1.
joachim99@52: 
joachim99@52:    If DELETED is 0 then LINE0 is the number of the line before
joachim99@52:    which the insertion was done; vice versa for INSERTED and LINE1.  */
joachim99@52: 
joachim99@52: struct change
joachim99@52: {
joachim99@52:   struct change *link;		/* Previous or next edit command  */
joachim99@52:   lin inserted;			/* # lines of file 1 changed here.  */
joachim99@52:   lin deleted;			/* # lines of file 0 changed here.  */
joachim99@52:   lin line0;			/* Line number of 1st deleted line.  */
joachim99@52:   lin line1;			/* Line number of 1st inserted line.  */
joachim99@52:   bool ignore;			/* Flag used in context.c.  */
joachim99@52: };
joachim99@52: 
joachim99@52: /* Structures that describe the input files.  */
joachim99@52: 
joachim99@52: /* Data on one input file being compared.  */
joachim99@52: 
joachim99@52: struct file_data {
joachim99@52: #if 0
joachim99@52:     int             desc;	/* File descriptor  */
joachim99@52:     char const      *name;	/* File name  */
joachim99@52:     struct stat     stat;	/* File status */
joachim99@52: #endif
joachim99@52:     /* Buffer in which text of file is read.  */
joachim99@52:     word *buffer;
joachim99@52: 
joachim99@52:     /* Allocated size of buffer, in bytes.  Always a multiple of
joachim99@52:        sizeof *buffer.  */
joachim99@52:     size_t bufsize;
joachim99@52: 
joachim99@52:     /* Number of valid bytes now in the buffer.  */
joachim99@52:     size_t buffered;
joachim99@52: 
joachim99@52:     /* Array of pointers to lines in the file.  */
joachim99@52:     char const **linbuf;
joachim99@52: 
joachim99@52:     /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines.
joachim99@52:        linebuf[linbuf_base ... buffered_lines - 1] are possibly differing.
joachim99@52:        linebuf[linbuf_base ... valid_lines - 1] contain valid data.
joachim99@52:        linebuf[linbuf_base ... alloc_lines - 1] are allocated.  */
joachim99@52:     lin linbuf_base, buffered_lines, valid_lines, alloc_lines;
joachim99@52: 
joachim99@52:     /* Pointer to end of prefix of this file to ignore when hashing.  */
joachim99@52:     char const *prefix_end;
joachim99@52: 
joachim99@52:     /* Count of lines in the prefix.
joachim99@52:        There are this many lines in the file before linbuf[0].  */
joachim99@52:     lin prefix_lines;
joachim99@52: 
joachim99@52:     /* Pointer to start of suffix of this file to ignore when hashing.  */
joachim99@52:     char const *suffix_begin;
joachim99@52: 
joachim99@52:     /* Vector, indexed by line number, containing an equivalence code for
joachim99@52:        each line.  It is this vector that is actually compared with that
joachim99@52:        of another file to generate differences.  */
joachim99@52:     lin *equivs;
joachim99@52: 
joachim99@52:     /* Vector, like the previous one except that
joachim99@52:        the elements for discarded lines have been squeezed out.  */
joachim99@52:     lin *undiscarded;
joachim99@52: 
joachim99@52:     /* Vector mapping virtual line numbers (not counting discarded lines)
joachim99@52:        to real ones (counting those lines).  Both are origin-0.  */
joachim99@52:     lin *realindexes;
joachim99@52: 
joachim99@52:     /* Total number of nondiscarded lines.  */
joachim99@52:     lin nondiscarded_lines;
joachim99@52: 
joachim99@52:     /* Vector, indexed by real origin-0 line number,
joachim99@52:        containing TRUE for a line that is an insertion or a deletion.
joachim99@52:        The results of comparison are stored here.  */
joachim99@52:     bool *changed;
joachim99@52: 
joachim99@52:     /* 1 if file ends in a line with no final newline.  */
joachim99@52:     bool missing_newline;
joachim99@52: 
joachim99@52:     /* 1 if at end of file.  */
joachim99@52:     bool eof;
joachim99@52: 
joachim99@52:     /* 1 more than the maximum equivalence value used for this or its
joachim99@52:        sibling file.  */
joachim99@52:     lin equiv_max;
joachim99@52: };
joachim99@52: 
joachim99@52: /* The file buffer, considered as an array of bytes rather than
joachim99@52:    as an array of words.  */
joachim99@52: #define FILE_BUFFER(f) ((char *) (f)->buffer)
joachim99@52: 
joachim99@52: /* Data on two input files being compared.  */
joachim99@52: 
joachim99@52: struct comparison
joachim99@52:   {
joachim99@52:     struct file_data file[2];
joachim99@52:     struct comparison const *parent;  /* parent, if a recursive comparison */
joachim99@52:   };
joachim99@52: 
joachim99@52: /* Describe the two files currently being compared.  */
joachim99@52: 
joachim99@53: struct file_data files[2];
joachim99@52: 
joachim99@52: /* Stdio stream to output diffs to.  */
joachim99@52: 
joachim99@53: FILE *outfile;
joachim99@52: 
joachim99@52: /* Declare various functions.  */
joachim99@52: 
joachim99@52: /* analyze.c */
joachim99@52: struct change* diff_2_files (struct comparison *);
joachim99@52: 
joachim99@52: /* context.c */
joachim99@52: void print_context_header (struct file_data[], bool);
joachim99@52: void print_context_script (struct change *, bool);
joachim99@52: 
joachim99@52: /* dir.c */
joachim99@52: int diff_dirs (struct comparison const *, int (*) (struct comparison const *, char const *, char const *));
joachim99@52: 
joachim99@52: /* ed.c */
joachim99@52: void print_ed_script (struct change *);
joachim99@52: void pr_forward_ed_script (struct change *);
joachim99@52: 
joachim99@52: /* ifdef.c */
joachim99@52: void print_ifdef_script (struct change *);
joachim99@52: 
joachim99@52: /* io.c */
joachim99@52: void file_block_read (struct file_data *, size_t);
joachim99@52: bool read_files (struct file_data[], bool);
joachim99@52: 
joachim99@52: /* normal.c */
joachim99@52: void print_normal_script (struct change *);
joachim99@52: 
joachim99@52: /* rcs.c */
joachim99@52: void print_rcs_script (struct change *);
joachim99@52: 
joachim99@52: /* side.c */
joachim99@52: void print_sdiff_script (struct change *);
joachim99@52: 
joachim99@52: /* util.c */
joachim99@53: //extern char const change_letter[4];
joachim99@53: //extern char const pr_program[];
joachim99@52: char *concat (char const *, char const *, char const *);
joachim99@52: char *dir_file_pathname (char const *, char const *);
joachim99@52: bool lines_differ (char const *, char const *);
joachim99@52: lin translate_line_number (struct file_data const *, lin);
joachim99@52: struct change *find_change (struct change *);
joachim99@52: struct change *find_reverse_change (struct change *);
joachim99@52: void *zalloc (size_t);
joachim99@52: enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
joachim99@52: void begin_output (void);
joachim99@52: void debug_script (struct change *);
joachim99@52: void fatal (char const *) __attribute__((noreturn));
joachim99@52: void finish_output (void);
joachim99@52: void message (char const *, char const *, char const *);
joachim99@52: void message5 (char const *, char const *, char const *, char const *, char const *);
joachim99@52: void output_1_line (char const *, char const *, char const *, char const *);
joachim99@52: void perror_with_name (char const *);
joachim99@52: void pfatal_with_name (char const *) __attribute__((noreturn));
joachim99@52: void print_1_line (char const *, char const * const *);
joachim99@52: void print_message_queue (void);
joachim99@52: void print_number_range (char, struct file_data *, lin, lin);
joachim99@52: void print_script (struct change *, struct change * (*) (struct change *), void (*) (struct change *));
joachim99@52: void setup_output (char const *, char const *, bool);
joachim99@52: void translate_range (struct file_data const *, lin, lin, long *, long *);
joachim99@52: 
joachim99@52: /* version.c */
joachim99@53: //extern char const version_string[];
joachim99@53: 
joachim99@53: private:
joachim99@53:    // gnudiff_analyze.cpp
joachim99@53:    lin diag (lin xoff, lin xlim, lin yoff, lin ylim, bool find_minimal, struct partition *part);
joachim99@53:    void compareseq (lin xoff, lin xlim, lin yoff, lin ylim, bool find_minimal);
joachim99@53:    void discard_confusing_lines (struct file_data filevec[]);
joachim99@53:    void shift_boundaries (struct file_data filevec[]);
joachim99@53:    struct change * add_change (lin line0, lin line1, lin deleted, lin inserted, struct change *old);
joachim99@53:    struct change * build_reverse_script (struct file_data const filevec[]);
joachim99@53:    struct change* build_script (struct file_data const filevec[]);
joachim99@53:    
joachim99@53:    // gnudiff_io.cpp
joachim99@53:    void find_and_hash_each_line (struct file_data *current);
joachim99@53:    void prepare_text (struct file_data *current);
joachim99@53:    void find_identical_ends (struct file_data filevec[]);
joachim99@53: 
joachim99@53:    // gnudiff_xmalloc.cpp
joachim99@53:    void *xmalloc (size_t n);
joachim99@53:    void *xcalloc (size_t n, size_t s);
joachim99@53:    void *xrealloc(void *p, size_t n);
joachim99@53:    char *xstrdup (const char *str);
joachim99@53:    void xalloc_die (void);
joachim99@53:    
joachim99@53:    inline bool isWhite( char c )
joachim99@53:    {
joachim99@53:       return c==' ' || c=='\t' ||  c=='\r';
joachim99@53:    }
joachim99@53: }; // class GnuDiff
joachim99@53: 
joachim99@53: # define XMALLOC(Type, N_items) ((Type *) xmalloc (sizeof (Type) * (N_items)))
joachim99@53: # define XCALLOC(Type, N_items) ((Type *) xcalloc (sizeof (Type), (N_items)))
joachim99@53: # define XREALLOC(Ptr, Type, N_items) \
joachim99@53:   ((Type *) xrealloc ((void *) (Ptr), sizeof (Type) * (N_items)))
joachim99@53: 
joachim99@53: /* Declare and alloc memory for VAR of type TYPE. */
joachim99@53: # define NEW(Type, Var)  Type *(Var) = XMALLOC (Type, 1)
joachim99@53: 
joachim99@53: /* Free VAR only if non NULL. */
joachim99@53: # define XFREE(Var)	\
joachim99@53:    do {                 \
joachim99@53:       if (Var)          \
joachim99@53:         free (Var);     \
joachim99@53:    } while (0)
joachim99@53: 
joachim99@53: /* Return a pointer to a malloc'ed copy of the array SRC of NUM elements. */
joachim99@53: # define CCLONE(Src, Num) \
joachim99@53:   (memcpy (xmalloc (sizeof (*Src) * (Num)), (Src), sizeof (*Src) * (Num)))
joachim99@53: 
joachim99@53: /* Return a malloc'ed copy of SRC. */
joachim99@53: # define CLONE(Src) CCLONE (Src, 1)
joachim99@53: 
joachim99@53: #endif