Skip long lines (fixes ggreer#189)

doc/ag.1

@@ -163,6 +163,12 @@ Recursively search for PATTERN in PATH\. Like grep or ack, but faster\.
 \~\~\~\~ Print matches on very long lines (> 2k characters by default)
 .
 .P
+\fB\-\-long\-line\-length\fR:
+.
+.br
+\~\~\~\~ Length limit for long lines (for \-\-print\-long\-lines)
+.
+.P
 \fB\-Q \-\-literal\fR:
 .
 .br

doc/ag.1.md

@@ -61,6 +61,8 @@ Recursively search for PATTERN in PATH. Like grep or ack, but faster.
     Use a pager such as less. Use `--nopager` to override. This option is also ignored if output is piped to another program.
   * `--print-long-lines`:
     Print matches on very long lines (> 2k characters by default)
+  * `--long-line-length`:
+    Length limit for long lines (for --print-long-lines)
   * `-Q --literal`:
     Do not parse PATTERN as a regular expression. Try to match it literally.
   * `-s --case-sensitive`:

src/options.c

@@ -78,6 +78,7 @@ Search options:\n\
 -p --path-to-agignore STRING\n\
                         Use .agignore file at STRING\n\
 --print-long-lines      Print matches on very long lines (Default: >2k characters)\n\
+--long-line-length      Length limit for long lines (for --print-long-lines)\n\
 -Q --literal            Don't parse PATTERN as a regular expression\n\
 -s --case-sensitive     Match case sensitively (Enabled by default)\n\
 -S --smart-case         Match case insensitively unless PATTERN contains\n\
@@ -117,6 +118,7 @@ void init_options() {
     opts.color_path = ag_strdup(color_path);
     opts.color_match = ag_strdup(color_match);
     opts.color_line_number = ag_strdup(color_line_number);
+    opts.long_line_length = 2000;
 }
 
 void cleanup_options() {
@@ -197,6 +199,7 @@ void parse_options(int argc, char **argv, char **base_paths[], char **paths[]) {
         { "invert-match", no_argument, &opts.invert_match, 1 },
         { "line-numbers", no_argument, &opts.print_line_numbers, 2 },
         { "literal", no_argument, NULL, 'Q' },
+        { "long-line-length", required_argument, &opts.long_line_length, 0},
         { "match", no_argument, &useless, 0 },
         { "max-count", required_argument, NULL, 'm' },
         { "no-numbers", no_argument, NULL, 0 },
@@ -382,6 +385,9 @@ void parse_options(int argc, char **argv, char **base_paths[], char **paths[]) {
                 } else if (strcmp(longopts[opt_index].name, "ignore") == 0) {
                     add_ignore_pattern(root_ignores, optarg);
                     break;
+                } else if (strcmp(longopts[opt_index].name, "long-line-length") == 0) {
+                    opts.long_line_length = atoi(optarg);
+                    break;
                 } else if (strcmp(longopts[opt_index].name, "nopager") == 0) {
                     out_fd = stdout;
                     opts.pager = NULL;

src/options.h

@@ -44,7 +44,8 @@ typedef struct {
     int print_filename_only;
     int print_heading;
     int print_line_numbers;
-    int print_long_lines; /* TODO: support this in print.c */
+    int print_long_lines;
+    int long_line_length;
     pcre *re;
     pcre_extra *re_extra;
     int recurse_dirs;

src/search.c

@@ -52,6 +52,17 @@ void search_buf(const char *buf, const int buf_len,
                 break;
             }
 
+            if (!opts.print_long_lines) {
+                int line_length = get_line_length(buf, buf_len, match_ptr - buf, opts.query_len + match_ptr - buf);
+                if (line_length >= opts.long_line_length) {
+                    log_debug("Too long line: %d > %d.\n", line_length, opts.long_line_length);
+                    const char *end = match_ptr + opts.query_len;
+                    buf_offset = end - buf;
+                    match_ptr += opts.query_len;
+                    continue;
+                }
+            }
+
             if (opts.word_regexp) {
                 const char *start = match_ptr;
                 const char *end = match_ptr + opts.query_len;
@@ -100,6 +111,14 @@ void search_buf(const char *buf, const int buf_len,
             log_debug("Regex match found. File %s, offset %i bytes.", dir_full_path, offset_vector[0]);
             buf_offset = offset_vector[1];
 
+            if (!opts.print_long_lines) {
+                int line_length = get_line_length(buf, buf_len, offset_vector[0], offset_vector[1]);
+                if (line_length >= opts.long_line_length) {
+                    log_debug("Too long line: %d > %d.\n", line_length, opts.long_line_length);
+                    continue;
+                }
+            }
+
             /* TODO: copy-pasted from above. FIXME */
             if ((size_t)matches_len + matches_spare >= matches_size) {
                 matches_size = matches ? matches_size * 2 : 100;
@@ -159,7 +178,9 @@ void search_stream(FILE *stream, const char *path) {
     size_t line_cap = 0;
 
     while ((line_len = getline(&line, &line_cap, stream)) > 0) {
-        search_buf(line, line_len, path);
+        if (opts.print_long_lines || line_len < opts.long_line_length) {
+            search_buf(line, line_len, path);
+        }
     }
 
     free(line);

src/util.c

@@ -470,3 +470,25 @@ int vasprintf(char **ret, const char *fmt, va_list args) {
     return rv;
 }
 #endif
+
+int get_line_length(const char* buf, const int buf_len,
+                    const int match_start, const int match_end) {
+    /* count the number of characters before and after the match */
+    int line_length = match_end - match_start;
+    const char* c;
+    for(c = buf + match_start - 1; c >= buf; --c) {
+        if (*c == '\n') {
+            break;
+        }
+        line_length ++;
+    }
+
+    for(c = buf + match_end; c < buf + buf_len; ++c) {
+        if (*c == '\n') {
+            break;
+        }
+        line_length ++;
+    }
+
+    return line_length;
+}

src/util.h

@@ -96,3 +96,6 @@ int vasprintf(char **ret, const char *fmt, va_list args);
 #endif
 
 #endif
+
+int get_line_length(const char* buf, const int buf_len,
+                    const int match_start, const int match_end);