Skip long lines (fixes ggreer#189)

doc/ag.1

@@ -159,6 +159,10 @@ When searching a stream, print all lines even if they don\'t match\.
 Do not parse PATTERN as a regular expression\. Try to match it literally\.
 .
 .TP
+\fB\-\-long\-line\-length\fR:
+Length limit for long lines (for \-\-print\-long\-lines)
+.
+.TP
 \fB\-s \-\-case\-sensitive\fR
 Match case\-sensitively\.
 .

doc/ag.1.md

@@ -110,7 +110,7 @@ Recursively search for PATTERN in PATH. Like grep or ack, but faster.
 
   * `-o --only-matching`:
     Print only the matching part of the lines.
-  
+
   * `-p --path-to-agignore STRING`:
     Provide a path to a specific .agignore file.
 
@@ -124,6 +124,9 @@ Recursively search for PATTERN in PATH. Like grep or ack, but faster.
   * `--passthrough`:
     When searching a stream, print all lines even if they don't match.
 
+  * `--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.
 

src/options.c

@@ -66,6 +66,7 @@ Output Options:\n\
                           when searching streams\n\
   -o --only-matching      Prints only the matching part of the lines\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\
      --passthrough        When searching a stream, print all lines even if they\n\
                           don't match\n\
      --silent             Suppress all log messages, including errors\n\
@@ -150,6 +151,7 @@ void init_options(void) {
     opts.color_match = ag_strdup(color_match);
     opts.color_line_number = ag_strdup(color_line_number);
     opts.use_thread_affinity = TRUE;
+    opts.long_line_length = 2000;
 }
 
 void cleanup_options(void) {
@@ -249,6 +251,7 @@ void parse_options(int argc, char **argv, char **base_paths[], char **paths[]) {
         { "line-numbers", no_argument, &opts.print_line_numbers, 2 },
         { "list-file-types", no_argument, &list_file_types, 1 },
         { "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-" is deprecated. Remove these eventually. */
@@ -490,6 +493,8 @@ void parse_options(int argc, char **argv, char **base_paths[], char **paths[]) {
                 } else if (strcmp(longopts[opt_index].name, "nofilename") == 0) {
                     opts.print_path = PATH_PRINT_NOTHING;
                     opts.print_line_numbers = FALSE;
+                } 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;

src/options.h

@@ -60,7 +60,8 @@ typedef struct {
     int print_filename_only;
     int print_path;
     int print_line_numbers;
-    int print_long_lines; /* TODO: support this in print.c */
+    int print_long_lines;
+    int long_line_length;
     int passthrough;
     pcre *re;
     pcre_extra *re_extra;

src/search.c

@@ -52,6 +52,17 @@ void search_buf(const char *buf, const size_t 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;
@@ -101,6 +112,14 @@ void search_buf(const char *buf, const size_t buf_len,
                 log_debug("Regex match is of length zero. Advancing offset one byte.");
             }
 
+            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 (matches_len + matches_spare >= matches_size) {
                 matches_size = matches ? matches_size * 2 : 100;
@@ -179,8 +198,10 @@ void search_stream(FILE *stream, const char *path) {
     size_t i;
 
     for (i = 1; (line_len = getline(&line, &line_cap, stream)) > 0; i++) {
-        opts.stream_line_num = i;
-        search_buf(line, line_len, path);
+        if (opts.print_long_lines || line_len < opts.long_line_length) {
+            opts.stream_line_num = i;
+            search_buf(line, line_len, path);
+        }
     }
 
     free(line);

src/util.c

@@ -637,3 +637,25 @@ 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

@@ -111,3 +111,6 @@ 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);