[bug-diffutils] bug#18641: [PATCH] diff: Add --only=REGEX option to parse only files that match REGEX

[Daniel Hilst Selli]

I run to this problem some time ago. I need to parse to trees, but
only the source files (*.[hcS]). I see some find solutions, but that
don't work for me since I have new files created on modified
tree. Best solution would be patch diff. So here I am.

This patches add a new option (--only=REGEX) that parse only files
that match such regexp. Multiple regexps can be used and they will be
OR'ed. Also, --exclude is processed before --only, so will can still
exclude files before they match REGEX.

This commit don't update man pages and info docs, I looking for it.

Cheers,
>8---------------------8<
* src/diff.h: XTERN only_regexp added.
* src/diff.c: global only_regexp_list, and ONLY_OPTION added.
              "only-that-matches" element added to longopts.
              Help string added to global option_help_msgid array.
* src/diff.c: (main) only_regexp_list.buf initalized.
              ONLY_OPITON processed on options loop
* src/dir.c:  (dir_read) Added an `if' to process --only option. Was
              added after --exclude processing. This way user can
              exclude files that still match --only REGEX.
---
 src/diff.c | 12 ++++++++++++
 src/diff.h |  3 +++
 src/dir.c  | 10 ++++++++++
 3 files changed, 25 insertions(+)

diff --git a/src/diff.c b/src/diff.c
index 397815e..003eb59 100644
--- a/src/diff.c
+++ b/src/diff.c
@@ -84,6 +84,9 @@ static struct regexp_list function_regexp_list;
 /* Ignore changes affecting only lines that match these regexps.  */
 static struct regexp_list ignore_regexp_list;
 
+/* Parse only files that match this regexp. */
+static struct regexp_list only_regexp_list;
+
 #if O_BINARY
 /* Use binary I/O when reading and writing data (--binary).
    On POSIX hosts, this has no effect.  */
@@ -126,6 +129,7 @@ enum
   SUPPRESS_COMMON_LINES_OPTION,
   TABSIZE_OPTION,
   TO_FILE_OPTION,
+  ONLY_OPTION,
 
   /* These options must be in sequence.  */
   UNCHANGED_LINE_FORMAT_OPTION,
@@ -191,6 +195,7 @@ static struct option const longopts[] =
   {"normal", 0, 0, NORMAL_OPTION},
   {"old-group-format", 1, 0, OLD_GROUP_FORMAT_OPTION},
   {"old-line-format", 1, 0, OLD_LINE_FORMAT_OPTION},
+  {"only-that-matches", 1, 0, ONLY_OPTION},
   {"paginate", 0, 0, 'l'},
   {"rcs", 0, 0, 'n'},
   {"recursive", 0, 0, 'r'},
@@ -282,6 +287,7 @@ main (int argc, char **argv)
   c_stack_action (0);
   function_regexp_list.buf = &function_regexp;
   ignore_regexp_list.buf = &ignore_regexp;
+  only_regexp_list.buf = &only_regexp;
   re_set_syntax (RE_SYNTAX_GREP | RE_NO_POSIX_BACKTRACKING);
   excluded = new_exclude ();
 
@@ -609,6 +615,10 @@ main (int argc, char **argv)
          specify_value (&to_file, optarg, "--to-file");
          break;
 
+        case ONLY_OPTION:
+          add_regexp (&only_regexp_list, optarg);
+          break;
+
        case UNCHANGED_LINE_FORMAT_OPTION:
        case OLD_LINE_FORMAT_OPTION:
        case NEW_LINE_FORMAT_OPTION:
@@ -696,6 +706,7 @@ main (int argc, char **argv)
 
   summarize_regexp_list (&function_regexp_list);
   summarize_regexp_list (&ignore_regexp_list);
+  summarize_regexp_list (&only_regexp_list);
 
   if (output_style == OUTPUT_IFDEF)
     {
@@ -883,6 +894,7 @@ static char const * const option_help_msgid[] = {
   N_("    --no-ignore-file-name-case  consider case when comparing file 
names"),
   N_("-x, --exclude=PAT               exclude files that match PAT"),
   N_("-X, --exclude-from=FILE         exclude files that match any pattern in 
FILE"),
+  N_("    --only=REGEX                only parse files that match REGEX"),
   N_("-S, --starting-file=FILE        start with FILE when comparing 
directories"),
   N_("    --from-file=FILE1           compare FILE1 to all operands;\n"
      "                                  FILE1 can be a directory"),
diff --git a/src/diff.h b/src/diff.h
index e9f0471..1df36f9 100644
--- a/src/diff.h
+++ b/src/diff.h
@@ -145,6 +145,9 @@ XTERN char *file_label[2];
 /* Regexp to identify function-header lines (-F).  */
 XTERN struct re_pattern_buffer function_regexp;
 
+/* Parse only file that match this regexp (--only) */
+XTERN struct re_pattern_buffer only_regexp;
+
 /* Ignore changes that affect only lines matching this regexp (-I).  */
 XTERN struct re_pattern_buffer ignore_regexp;
 
diff --git a/src/dir.c b/src/dir.c
index d3b0a2d..5fcd06f 100644
--- a/src/dir.c
+++ b/src/dir.c
@@ -99,6 +99,16 @@ dir_read (struct file_data const *dir, struct dirdata 
*dirdata)
          if (excluded_file_name (excluded, d_name))
            continue;
 
+          if (only_regexp.fastmap)
+            {
+              int only_match = re_search (&only_regexp, d_name, d_size, 0, 
d_size, NULL);
+
+              if (-1 == only_match)
+                continue;
+              else if (-2 == only_match)
+                fatal("Error while matching --only option");
+            }
+
          while (data_alloc < data_used + d_size)
            {
              if (PTRDIFF_MAX / 2 <= data_alloc)
-- 
2.1.2