From 483a454c8dfaf7c5eaa9a514ad6728431250189b Mon Sep 17 00:00:00 2001
From: simon <me@simon987.net>
Date: Sat, 22 Feb 2020 16:55:35 -0500
Subject: [PATCH] --exclude argument #26

---
 CMakeLists.txt |  2 ++
 src/cli.c      | 21 +++++++++++++++++++++
 src/cli.h      |  6 ++++++
 src/ctx.h      |  2 ++
 src/io/walk.c  | 12 +++++++++++-
 src/main.c     |  8 +++-----
 src/sist.h     |  1 +
 7 files changed, 46 insertions(+), 6 deletions(-)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 80b03c2..46900dc 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -137,6 +137,8 @@ TARGET_LINK_LIBRARIES(
         ${PROJECT_SOURCE_DIR}/lib/libcrypto.a
         ${PROJECT_SOURCE_DIR}/lib/libssl.a
         dl
+
+        pcre
 )
 
 add_custom_target(
diff --git a/src/cli.c b/src/cli.c
index d0663df..4a53198 100644
--- a/src/cli.c
+++ b/src/cli.c
@@ -162,6 +162,26 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
         args->tesseract_path = path;
     }
 
+    if (args->exclude_regex != NULL) {
+        const char *error;
+        int error_offset;
+
+        pcre *re = pcre_compile(args->exclude_regex, 0, &error, &error_offset, 0);
+        if (error != NULL) {
+            LOG_FATALF("cli.c", "pcre_compile returned error: %s (offset:%d)", error, error_offset)
+        }
+
+        pcre_extra *re_extra = pcre_study(re, 0, &error);
+        if (error != NULL) {
+            LOG_FATALF("cli.c", "pcre_study returned error: %s", error)
+        }
+
+        ScanCtx.exclude = re;
+        ScanCtx.exclude_extra = re_extra;
+    } else {
+        ScanCtx.exclude = NULL;
+    }
+
     LOG_DEBUGF("cli.c", "arg quality=%f", args->quality)
     LOG_DEBUGF("cli.c", "arg size=%d", args->size)
     LOG_DEBUGF("cli.c", "arg content_size=%d", args->content_size)
@@ -175,6 +195,7 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
     LOG_DEBUGF("cli.c", "arg archive=%s", args->archive)
     LOG_DEBUGF("cli.c", "arg tesseract_lang=%s", args->tesseract_lang)
     LOG_DEBUGF("cli.c", "arg tesseract_path=%s", args->tesseract_path)
+    LOG_DEBUGF("cli.c", "arg exclude=%s", args->exclude_regex)
 
     return 0;
 }
diff --git a/src/cli.h b/src/cli.h
index f0eb163..0bfecf3 100644
--- a/src/cli.h
+++ b/src/cli.h
@@ -18,10 +18,13 @@ typedef struct scan_args {
     archive_mode_t archive_mode;
     char *tesseract_lang;
     const char *tesseract_path;
+    char *exclude_regex;
 } scan_args_t;
 
 scan_args_t *scan_args_create();
+
 void scan_args_destroy(scan_args_t *args);
+
 int scan_args_validate(scan_args_t *args, int argc, const char **argv);
 
 typedef struct index_args {
@@ -45,12 +48,15 @@ typedef struct web_args {
 } web_args_t;
 
 index_args_t *index_args_create();
+
 void index_args_destroy(index_args_t *args);
 
 web_args_t *web_args_create();
+
 void web_args_destroy(web_args_t *args);
 
 int index_args_validate(index_args_t *args, int argc, const char **argv);
+
 int web_args_validate(web_args_t *args, int argc, const char **argv);
 
 #endif
diff --git a/src/ctx.h b/src/ctx.h
index fc08744..f0da63b 100644
--- a/src/ctx.h
+++ b/src/ctx.h
@@ -29,6 +29,8 @@ struct {
     pthread_mutex_t mupdf_mu;
     char * tesseract_lang;
     const char * tesseract_path;
+    pcre *exclude;
+    pcre_extra *exclude_extra;
 } ScanCtx;
 
 struct {
diff --git a/src/io/walk.c b/src/io/walk.c
index d1debc4..afd1cec 100644
--- a/src/io/walk.c
+++ b/src/io/walk.c
@@ -28,8 +28,18 @@ parse_job_t *create_fs_parse_job(const char *filepath, const struct stat *info,
     return job;
 }
 
+int sub_strings[30];
+#define EXCLUDED(str) (pcre_exec(ScanCtx.exclude, ScanCtx.exclude_extra, filepath, strlen(filepath), 0, 0, sub_strings, sizeof(sub_strings)) >= 0)
+
 int handle_entry(const char *filepath, const struct stat *info, int typeflag, struct FTW *ftw) {
-    if (ftw->level <= ScanCtx.depth && typeflag == FTW_F && S_ISREG(info->st_mode)) {
+
+    if (typeflag == FTW_F && S_ISREG(info->st_mode) && ftw->level <= ScanCtx.depth) {
+
+        if (ScanCtx.exclude != NULL && EXCLUDED(filepath)) {
+            LOG_DEBUGF("walk.c", "Excluded: %s", filepath)
+            return 0;
+        }
+
         parse_job_t *job = create_fs_parse_job(filepath, info, ftw->base);
         tpool_add_work(ScanCtx.pool, parse, job);
     }
diff --git a/src/main.c b/src/main.c
index 7a0760a..7a63369 100644
--- a/src/main.c
+++ b/src/main.c
@@ -240,6 +240,7 @@ int main(int argc, const char *argv[]) {
                                                           "shallow: Don't parse archives inside archives. DEFAULT: recurse"),
             OPT_STRING(0, "ocr", &scan_args->tesseract_lang, "Tesseract language (use tesseract --list-langs to see "
                                                              "which are installed on your machine)"),
+            OPT_STRING('e', "exclude", &scan_args->exclude_regex, "Files that match this regex will not be scanned"),
 
             OPT_GROUP("Index options"),
             OPT_STRING(0, "es-url", &common_es_url, "Elasticsearch url with port. DEFAULT=http://localhost:9200"),
@@ -286,9 +287,7 @@ int main(int argc, const char *argv[]) {
         }
         sist2_scan(scan_args);
 
-    }
-
-    else if (strcmp(argv[0], "index") == 0) {
+    } else if (strcmp(argv[0], "index") == 0) {
 
         int err = index_args_validate(index_args, argc, argv);
         if (err != 0) {
@@ -304,8 +303,7 @@ int main(int argc, const char *argv[]) {
         }
         sist2_web(web_args);
 
-    }
-    else {
+    } else {
         fprintf(stderr, "Invalid command: '%s'\n", argv[0]);
         argparse_usage(&argparse);
         return 1;
diff --git a/src/sist.h b/src/sist.h
index 79d1662..dc4209e 100644
--- a/src/sist.h
+++ b/src/sist.h
@@ -35,6 +35,7 @@
 #include <libxml/xmlstring.h>
 #define BOOL int
 #include <tesseract/capi.h>
+#include <pcre.h>
 
 #include <onion/onion.h>
 #include <onion/handler.h>