Version bump

Fix link to scripting.

.github/ISSUE_TEMPLATE/issue-template.md

@@ -9,7 +9,7 @@ assignees: ''
 
 sist2 version:
 
-Platform (Linux or Docker):
+Platform (Linux or Docker, x86-64 or arm64):
 
 Elasticsearch version:
 

DockerArm64/Dockerfile

@@ -0,0 +1,22 @@
+FROM ubuntu:19.10
+MAINTAINER simon987 <me@simon987.net>
+
+RUN apt update
+RUN apt install -y libglib2.0-0 libcurl4 libmagic1 libharfbuzz-bin libopenjp2-7 libarchive13 liblzma5 libzstd1 liblz4-1 \
+ curl libtiff5 libpng16-16 libpcre3
+
+RUN mkdir -p /usr/share/tessdata && \
+    cd /usr/share/tessdata/ && \
+    curl -o /usr/share/tessdata/hin.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/hin.traineddata &&\
+    curl -o /usr/share/tessdata/jpn.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/jpn.traineddata &&\
+    curl -o /usr/share/tessdata/eng.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/eng.traineddata &&\
+    curl -o /usr/share/tessdata/fra.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/fra.traineddata &&\
+    curl -o /usr/share/tessdata/rus.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/rus.traineddata &&\
+    curl -o /usr/share/tessdata/spa.traineddata https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/spa.traineddata && ls -lh
+
+ADD sist2_arm64 /root/sist2
+
+ENV LANG C.UTF-8
+ENV LC_ALL C.UTF-8
+
+ENTRYPOINT ["/root/sist2"]

DockerArm64/build.sh

@@ -0,0 +1,13 @@
+rm ./sist2_arm64
+cp ../sist2_arm64.gz .
+gzip -d sist2_arm64.gz
+
+version=$(./sist2_arm64 --version)
+
+echo "Version ${version}"
+docker build . -t simon987/sist2-arm64:"${version}" -t simon987/sist2-arm64:latest
+
+docker push simon987/sist2-arm64:"${version}"
+docker push simon987/sist2-arm64:latest
+
+docker run --rm simon987/sist2-arm64 -v

README.md

@@ -2,6 +2,8 @@
 [![CodeFactor](https://www.codefactor.io/repository/github/simon987/sist2/badge?s=05daa325188aac4eae32c786f3d9cf4e0593f822)](https://www.codefactor.io/repository/github/simon987/sist2)
 [![Development snapshots](https://ci.simon987.net/app/rest/builds/buildType(Sist2_Build)/statusIcon)](https://files.simon987.net/artifacts/Sist2/Build/)
 
+**Demo**: [sist2.simon987.net](https://sist2.simon987.net/)
+
 # sist2
 
 sist2 (Simple incremental search tool)
@@ -124,7 +126,7 @@ binaries (GCC 7+ required).
 1. Install compile-time dependencies
 
    ```bash
-   vcpkg install lmdb cjson glib libarchive[core,bzip2,libxml2,lz4,lzma,lzo] pthread tesseract libxml2 ffmpeg zstd gtest mongoose libuuid libmagic libraw
+   vcpkg install lmdb cjson glib libarchive[core,bzip2,libxml2,lz4,lzma,lzo] pthread tesseract libxml2 ffmpeg zstd gtest mongoose libuuid libmagic libraw curl[core,ssl] jbig2dec brotli libmupdf
    ```
 
 2. Build

ci/build_arm64.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+VCPKG_ROOT="/vcpkg"
+
+rm *.gz
+
+rm -rf CMakeFiles CMakeCache.txt
+cmake -DSIST_DEBUG=off -DCMAKE_TOOLCHAIN_FILE="${VCPKG_ROOT}/scripts/buildsystems/vcpkg.cmake" .
+make -j 4
+strip sist2
+mv sist2 sist2_arm64
+gzip -9 sist2_arm64

docs/USAGE.md

@@ -49,19 +49,25 @@ Scan options
 Index options
     -t, --threads=<int>           Number of threads. DEFAULT=1
     --es-url=<str>                Elasticsearch url with port. DEFAULT=http://localhost:9200
+    --es-index=<str>              Elasticsearch index name. DEFAULT=sist2
     -p, --print                   Just print JSON documents to stdout.
     --script-file=<str>           Path to user script.
+    --async-script                Execute user script asynchronously.
     --batch-size=<int>            Index batch size. DEFAULT: 100
     -f, --force-reset             Reset Elasticsearch mappings and settings. (You must use this option the first time you use the index command)
 
 Web options
     --es-url=<str>                Elasticsearch url. DEFAULT=http://localhost:9200
+    --es-index=<str>              Elasticsearch index name. DEFAULT=sist2
     --bind=<str>                  Listen on this address. DEFAULT=localhost:4090
     --auth=<str>                  Basic auth in user:password format
     --tag-auth=<str>              Basic auth in user:password format for tagging
 
 Exec-script options
+    --es-url=<str>                Elasticsearch url. DEFAULT=http://localhost:9200
+    --es-index=<str>              Elasticsearch index name. DEFAULT=sist2
     --script-file=<str>           Path to user script.
+    --async-script                Execute user script asynchronously.
 Made by simon987 <me@simon987.net>. Released under GPL-3.0
 ```
 
@@ -239,10 +245,15 @@ it is currently unsupported and has no guaranties of back/forward compatibility.
  * `--es-url` 
  Elasticsearch url and port. If you are using docker, make sure that both containers are on the
  same network.
+ * `--es-index` 
+    Elasticsearch index name. DEFAULT=sist2
  * `-p, --print` 
     Print index in JSON format to stdout.
  * `--script-file` 
     Path to user script. See [Scripting](scripting.md).
+ * `--async-script` 
+    Use `wait_for_completion=false` elasticsearch option while executing user script.
+     (See [Elasticsearch documentation](https://www.elastic.co/guide/en/elasticsearch/reference/current/tasks.html))
  * `--batch-size=<int>` 
     Index batch size. Indexing is generally faster with larger batches, but payloads that
     are too large will fail and additional overhead for retrying with smaller sizes may slow
@@ -273,6 +284,8 @@ sist2 index --print ./my_index/ | jq | less
 
 ### Web options
  * `--es-url=<str>` Elasticsearch url.
+ * `--es-index` 
+    Elasticsearch index name. DEFAULT=sist2
  * `--bind=<str>` Listen on this address.
  * `--auth=<str>` Basic auth in user:password format
  * `--tag-auth=<str>` Basic auth in user:password format. Works the same way as the 
@@ -335,4 +348,4 @@ See [Automatic tagging](#automatic-tagging) for information about tag
 
 ### Automatic tagging
 
-See [scripting](docs/scripting.md) documentation.
+See [scripting](scripting.md) documentation.

schema/mappings.json

@@ -54,6 +54,10 @@
       "type": "integer",
       "index": false
     },
+    "pages": {
+      "type": "integer",
+      "index": false
+    },
     "mtime": {
       "type": "integer"
     },

scripts/mime.csv

@@ -347,7 +347,8 @@ text/javascript, js
 text/mcf, mcf
 text/pascal, pas
 text/PGP,
-text/plain, com|cmd|conf|def|g|idc|list|lst|mar|sdml|text|txt|md|groovy|license|properties|desktop|ini|rst|cmake|ipynb|readme|less|lo|go|yml|d|cs|hpp|srt|nfo|sfv|m3u|csv|eml
+text/plain, com|cmd|conf|def|g|idc|list|lst|mar|sdml|text|txt|md|groovy|license|properties|desktop|ini|rst|cmake|ipynb|readme|less|lo|go|yml|d|cs|hpp|srt|nfo|sfv|m3u|csv|eml|make|log|markdown|yaml
+application/vnd.coffeescript, coffee
 text/richtext, rt|rtf|rtx
 text/rtf,
 text/scriplet, wsc

src/cli.c

@@ -9,6 +9,7 @@
 #define DEFAULT_REWRITE_URL ""
 
 #define DEFAULT_ES_URL "http://localhost:9200"
+#define DEFAULT_ES_INDEX "sist2"
 #define DEFAULT_BATCH_SIZE 100
 
 #define DEFAULT_LISTEN_ADDRESS "localhost:4090"
@@ -135,6 +136,10 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
 
     if (args->name == NULL) {
         args->name = g_path_get_basename(args->output);
+    } else {
+        char* tmp = malloc(strlen(args->name) + 1);
+        strcpy(tmp, args->name);
+        args->name = tmp;
     }
 
     if (args->rewrite_url == NULL) {
@@ -283,6 +288,10 @@ int index_args_validate(index_args_t *args, int argc, const char **argv) {
         args->es_url = DEFAULT_ES_URL;
     }
 
+    if (args->es_index == NULL) {
+        args->es_index = DEFAULT_ES_INDEX;
+    }
+
     if (args->script_path != NULL) {
         if (load_script(args->script_path, &args->script) != 0) {
             return 1;
@@ -294,8 +303,10 @@ int index_args_validate(index_args_t *args, int argc, const char **argv) {
     }
 
     LOG_DEBUGF("cli.c", "arg es_url=%s", args->es_url)
+    LOG_DEBUGF("cli.c", "arg es_index=%s", args->es_index)
     LOG_DEBUGF("cli.c", "arg index_path=%s", args->index_path)
     LOG_DEBUGF("cli.c", "arg script_path=%s", args->script_path)
+    LOG_DEBUGF("cli.c", "arg async_script=%s", args->async_script)
     LOG_DEBUGF("cli.c", "arg script=%s", args->script)
     LOG_DEBUGF("cli.c", "arg print=%d", args->print)
     LOG_DEBUGF("cli.c", "arg batch_size=%d", args->batch_size)
@@ -321,6 +332,10 @@ int web_args_validate(web_args_t *args, int argc, const char **argv) {
         args->listen_address = DEFAULT_LISTEN_ADDRESS;
     }
 
+    if (args->es_index == NULL) {
+        args->es_index = DEFAULT_ES_INDEX;
+    }
+
     if (args->credentials != NULL) {
         char *ptr = strstr(args->credentials, ":");
         if (ptr == NULL) {
@@ -378,6 +393,7 @@ int web_args_validate(web_args_t *args, int argc, const char **argv) {
     }
 
     LOG_DEBUGF("cli.c", "arg es_url=%s", args->es_url)
+    LOG_DEBUGF("cli.c", "arg es_index=%s", args->es_index)
     LOG_DEBUGF("cli.c", "arg listen=%s", args->listen_address)
     LOG_DEBUGF("cli.c", "arg credentials=%s", args->credentials)
     LOG_DEBUGF("cli.c", "arg tag_credentials=%s", args->tag_credentials)
@@ -421,6 +437,10 @@ int exec_args_validate(exec_args_t *args, int argc, const char **argv) {
         args->es_url = DEFAULT_ES_URL;
     }
 
+    if (args->es_index == NULL) {
+        args->es_index = DEFAULT_ES_INDEX;
+    }
+
     if (args->script_path == NULL) {
         LOG_FATAL("cli.c", "--script-file argument is required");
     }

src/cli.h

@@ -35,17 +35,20 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv);
 
 typedef struct index_args {
     char *es_url;
+    char *es_index;
     const char *index_path;
     const char *script_path;
     char *script;
     int print;
     int batch_size;
+    int async_script;
     int force_reset;
     int threads;
 } index_args_t;
 
 typedef struct web_args {
     char *es_url;
+    char *es_index;
     char *listen_address;
     char *credentials;
     char *tag_credentials;
@@ -59,8 +62,10 @@ typedef struct web_args {
 
 typedef struct exec_args {
     char *es_url;
+    char *es_index;
     const char *index_path;
     const char *script_path;
+    int async_script;
     char *script;
 } exec_args_t;
 

src/ctx.h

@@ -58,6 +58,7 @@ typedef struct {
 
 typedef struct {
     char *es_url;
+    char *es_index;
     int batch_size;
     tpool_t *pool;
     store_t *tag_store;
@@ -66,6 +67,7 @@ typedef struct {
 
 typedef struct {
     char *es_url;
+    char *es_index;
     int index_count;
     char *auth_user;
     char *auth_pass;

src/index/elastic.c

@@ -9,6 +9,7 @@
 typedef struct es_indexer {
     int queued;
     char *es_url;
+    char *es_index;
     es_bulk_line_t *line_head;
     es_bulk_line_t *line_tail;
 } es_indexer_t;
@@ -17,11 +18,13 @@ typedef struct es_indexer {
 static __thread es_indexer_t *Indexer;
 
 void delete_queue(int max);
+
 void elastic_flush();
 
 void elastic_cleanup() {
     elastic_flush();
     if (Indexer != NULL) {
+        free(Indexer->es_index);
         free(Indexer->es_url);
         free(Indexer);
     }
@@ -32,7 +35,7 @@ void print_json(cJSON *document, const char uuid_str[UUID_STR_LEN]) {
     cJSON *line = cJSON_CreateObject();
 
     cJSON_AddStringToObject(line, "_id", uuid_str);
-    cJSON_AddStringToObject(line, "_index", "sist2");
+    cJSON_AddStringToObject(line, "_index", IndexCtx.es_index);
     cJSON_AddStringToObject(line, "_type", "_doc");
     cJSON_AddItemReferenceToObject(line, "_source", document);
 
@@ -64,10 +67,10 @@ void index_json(cJSON *document, const char uuid_str[UUID_STR_LEN]) {
     tpool_add_work(IndexCtx.pool, index_json_func, bulk_line);
 }
 
-void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]) {
+void execute_update_script(const char *script, int async, const char index_id[UUID_STR_LEN]) {
 
     if (Indexer == NULL) {
-        Indexer = create_indexer(IndexCtx.es_url);
+        Indexer = create_indexer(IndexCtx.es_url, IndexCtx.es_index);
     }
 
     cJSON *body = cJSON_CreateObject();
@@ -82,9 +85,16 @@ void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]
     char *str = cJSON_Print(body);
 
     char bulk_url[4096];
-    snprintf(bulk_url, 4096, "%s/sist2/_update_by_query?wait_for_completion=false", Indexer->es_url);
+    if (async) {
+        snprintf(bulk_url, sizeof(bulk_url), "%s/%s/_update_by_query?wait_for_completion=false", Indexer->es_url,
+                 Indexer->es_index);
+    } else {
+        snprintf(bulk_url, sizeof(bulk_url), "%s/%s/_update_by_query", Indexer->es_url, Indexer->es_index);
+    }
     response_t *r = web_post(bulk_url, str);
-    LOG_INFOF("elastic.c", "Executed user script <%d>", r->status_code);
+    if (!async) {
+        LOG_INFOF("elastic.c", "Executed user script <%d>", r->status_code);
+    }
     cJSON *resp = cJSON_Parse(r->body);
 
     cJSON_free(str);
@@ -99,11 +109,14 @@ void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]
         cJSON_free(error_str);
     }
 
+    if (async) {
+        cJSON *task = cJSON_GetObjectItem(resp, "task");
+        LOG_INFOF("elastic.c", "User script queued: %s/_tasks/%s", Indexer->es_url, task->valuestring);
+    }
+
     cJSON_Delete(resp);
 }
 
-#define ACTION_STR_LEN 91
-
 void *create_bulk_buffer(int max, int *count, size_t *buf_len) {
     es_bulk_line_t *line = Indexer->line_head;
     *count = 0;
@@ -115,20 +128,24 @@ void *create_bulk_buffer(int max, int *count, size_t *buf_len) {
 
     while (line != NULL && *count < max) {
         char action_str[256];
-        snprintf(action_str, 256,
+        snprintf(
-                 "{\"index\":{\"_id\":\"%s\", \"_type\":\"_doc\", \"_index\":\"sist2\"}}\n", line->uuid_str);
+                action_str, 256,
+                "{\"index\":{\"_id\":\"%s\",\"_type\":\"_doc\",\"_index\":\"%s\"}}\n",
+                line->uuid_str, Indexer->es_index
+        );
 
+        size_t action_str_len = strlen(action_str);
         size_t line_len = strlen(line->line);
 
-        while (buf_size + line_len + ACTION_STR_LEN > buf_capacity) {
+        while (buf_size + line_len + action_str_len > buf_capacity) {
             buf_capacity *= 2;
             buf = realloc(buf, buf_capacity);
         }
 
-        buf_size += line_len + ACTION_STR_LEN;
+        buf_size += line_len + action_str_len;
 
-        memcpy(buf + buf_cur, action_str, ACTION_STR_LEN);
+        memcpy(buf + buf_cur, action_str, action_str_len);
-        buf_cur += ACTION_STR_LEN;
+        buf_cur += action_str_len;
         memcpy(buf + buf_cur, line->line, line_len);
         buf_cur += line_len;
 
@@ -166,6 +183,21 @@ void print_errors(response_t *r) {
     free(tmp);
 }
 
+void print_error(response_t *r) {
+    char *tmp = malloc(r->size + 1);
+    memcpy(tmp, r->body, r->size);
+    *(tmp + r->size) = '\0';
+
+    cJSON *ret_json = cJSON_Parse(tmp);
+    if (cJSON_GetObjectItem(ret_json, "error") != NULL) {
+        char *str = cJSON_Print(cJSON_GetObjectItem(ret_json, "error"));
+        LOG_ERRORF("elastic.c", "%s\n", str);
+        cJSON_free(str);
+    }
+    cJSON_Delete(ret_json);
+    free(tmp);
+}
+
 void _elastic_flush(int max) {
 
     if (max == 0) {
@@ -178,7 +210,7 @@ void _elastic_flush(int max) {
     void *buf = create_bulk_buffer(max, &count, &buf_len);
 
     char bulk_url[4096];
-    snprintf(bulk_url, 4096, "%s/sist2/_bulk?pipeline=tie", Indexer->es_url);
+    snprintf(bulk_url, sizeof(bulk_url), "%s/%s/_bulk?pipeline=tie", Indexer->es_url, Indexer->es_index);
     response_t *r = web_post(bulk_url, buf);
 
     if (r->status_code == 0) {
@@ -248,7 +280,7 @@ void delete_queue(int max) {
 void elastic_flush() {
 
     if (Indexer == NULL) {
-        Indexer = create_indexer(IndexCtx.es_url);
+        Indexer = create_indexer(IndexCtx.es_url, IndexCtx.es_index);
     }
 
     _elastic_flush(Indexer->queued);
@@ -257,7 +289,7 @@ void elastic_flush() {
 void elastic_index_line(es_bulk_line_t *line) {
 
     if (Indexer == NULL) {
-        Indexer = create_indexer(IndexCtx.es_url);
+        Indexer = create_indexer(IndexCtx.es_url, IndexCtx.es_index);
     }
 
     if (Indexer->line_head == NULL) {
@@ -275,14 +307,18 @@ void elastic_index_line(es_bulk_line_t *line) {
     }
 }
 
-es_indexer_t *create_indexer(const char *url) {
+es_indexer_t *create_indexer(const char *url, const char *index) {
 
     char *es_url = malloc(strlen(url) + 1);
     strcpy(es_url, url);
 
+    char *es_index = malloc(strlen(index) + 1);
+    strcpy(es_index, index);
+
     es_indexer_t *indexer = malloc(sizeof(es_indexer_t));
 
     indexer->es_url = es_url;
+    indexer->es_index = es_index;
     indexer->queued = 0;
     indexer->line_head = NULL;
     indexer->line_tail = NULL;
@@ -290,31 +326,31 @@ es_indexer_t *create_indexer(const char *url) {
     return indexer;
 }
 
-void finish_indexer(char *script, char *index_id) {
+void finish_indexer(char *script, int async_script, char *index_id) {
 
     char url[4096];
 
-    snprintf(url, sizeof(url), "%s/sist2/_refresh", IndexCtx.es_url);
+    snprintf(url, sizeof(url), "%s/%s/_refresh", IndexCtx.es_url, IndexCtx.es_index);
     response_t *r = web_post(url, "");
     LOG_INFOF("elastic.c", "Refresh index <%d>", r->status_code);
     free_response(r);
 
     if (script != NULL) {
-        execute_update_script(script, index_id);
+        execute_update_script(script, async_script, index_id);
         free(script);
 
-        snprintf(url, sizeof(url), "%s/sist2/_refresh", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s/_refresh", IndexCtx.es_url, IndexCtx.es_index);
         r = web_post(url, "");
         LOG_INFOF("elastic.c", "Refresh index <%d>", r->status_code);
         free_response(r);
     }
 
-    snprintf(url, sizeof(url), "%s/sist2/_forcemerge", IndexCtx.es_url);
+    snprintf(url, sizeof(url), "%s/%s/_forcemerge", IndexCtx.es_url, IndexCtx.es_index);
     r = web_post(url, "");
     LOG_INFOF("elastic.c", "Merge index <%d>", r->status_code);
     free_response(r);
 
-    snprintf(url, sizeof(url), "%s/sist2/_settings", IndexCtx.es_url);
+    snprintf(url, sizeof(url), "%s/%s/_settings", IndexCtx.es_url, IndexCtx.es_index);
     r = web_put(url, "{\"index\":{\"refresh_interval\":\"1s\"}}");
     LOG_INFOF("elastic.c", "Set refresh interval <%d>", r->status_code);
     free_response(r);
@@ -324,7 +360,7 @@ void elastic_init(int force_reset) {
 
     // Check if index exists
     char url[4096];
-    snprintf(url, 4096, "%s/sist2", IndexCtx.es_url);
+    snprintf(url, sizeof(url), "%s/%s", IndexCtx.es_url, IndexCtx.es_index);
     response_t *r = web_get(url, 30);
     int index_exists = r->status_code == 200;
     free_response(r);
@@ -334,32 +370,38 @@ void elastic_init(int force_reset) {
         LOG_INFOF("elastic.c", "Delete index <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/sist2", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s", IndexCtx.es_url, IndexCtx.es_index);
         r = web_put(url, "");
+
+        if (r->status_code != 200) {
+            print_error(r);
+            LOG_FATAL("elastic.c", "Could not create index")
+        }
+
         LOG_INFOF("elastic.c", "Create index <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/sist2/_close", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s/_close", IndexCtx.es_url, IndexCtx.es_index);
         r = web_post(url, "");
         LOG_INFOF("elastic.c", "Close index <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/_ingest/pipeline/tie", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/_ingest/pipeline/tie", IndexCtx.es_url);
         r = web_put(url, pipeline_json);
         LOG_INFOF("elastic.c", "Create pipeline <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/sist2/_settings", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s/_settings", IndexCtx.es_url, IndexCtx.es_index);
         r = web_put(url, settings_json);
         LOG_INFOF("elastic.c", "Update settings <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/sist2/_mappings/_doc?include_type_name=true", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s/_mappings/_doc?include_type_name=true", IndexCtx.es_url, IndexCtx.es_index);
         r = web_put(url, mappings_json);
         LOG_INFOF("elastic.c", "Update mappings <%d>", r->status_code);
         free_response(r);
 
-        snprintf(url, 4096, "%s/sist2/_open", IndexCtx.es_url);
+        snprintf(url, sizeof(url), "%s/%s/_open", IndexCtx.es_url, IndexCtx.es_index);
         r = web_post(url, "");
         LOG_INFOF("elastic.c", "Open index <%d>", r->status_code);
         free_response(r);
@@ -368,12 +410,16 @@ void elastic_init(int force_reset) {
 
 cJSON *elastic_get_document(const char *uuid_str) {
     char url[4096];
-    snprintf(url, 4096, "%s/sist2/_doc/%s", WebCtx.es_url, uuid_str);
+    snprintf(url, sizeof(url), "%s/%s/_doc/%s", WebCtx.es_url, WebCtx.es_index, uuid_str);
 
     response_t *r = web_get(url, 3);
     cJSON *json = NULL;
     if (r->status_code == 200) {
-        json = cJSON_Parse(r->body);
+        char *tmp = malloc(r->size + 1);
+        memcpy(tmp, r->body, r->size);
+        *(tmp + r->size) = '\0';
+        json = cJSON_Parse(tmp);
+        free(tmp);
     }
     free_response(r);
     return json;
@@ -381,8 +427,8 @@ cJSON *elastic_get_document(const char *uuid_str) {
 
 char *elastic_get_status() {
     char url[4096];
-    snprintf(url, 4096,
+    snprintf(url, sizeof(url),
-             "%s/_cluster/state/metadata/sist2?filter_path=metadata.indices.*.state", WebCtx.es_url);
+             "%s/_cluster/state/metadata/%s?filter_path=metadata.indices.*.state", WebCtx.es_url, WebCtx.es_index);
 
     response_t *r = web_get(url, 30);
     cJSON *json = NULL;
@@ -390,12 +436,16 @@ char *elastic_get_status() {
     status[0] = '\0';
 
     if (r->status_code == 200) {
-        json = cJSON_Parse(r->body);
+        char *tmp = malloc(r->size + 1);
+        memcpy(tmp, r->body, r->size);
+        *(tmp + r->size) = '\0';
+        json = cJSON_Parse(tmp);
+        free(tmp);
         const cJSON *metadata = cJSON_GetObjectItem(json, "metadata");
         if (metadata != NULL) {
             const cJSON *indices = cJSON_GetObjectItem(metadata, "indices");
-            const cJSON *sist2 = cJSON_GetObjectItem(indices, "sist2");
+            const cJSON *index = cJSON_GetObjectItem(indices, WebCtx.es_index);
-            const cJSON *state = cJSON_GetObjectItem(sist2, "state");
+            const cJSON *state = cJSON_GetObjectItem(index, "state");
             strcpy(status, state->valuestring);
         }
     }

src/index/elastic.h

@@ -20,10 +20,10 @@ void print_json(cJSON *document, const char uuid_str[UUID_STR_LEN]);
 
 void index_json(cJSON *document, const char uuid_str[UUID_STR_LEN]);
 
-es_indexer_t *create_indexer(const char* es_url);
+es_indexer_t *create_indexer(const char *url, const char *index);
 
 void elastic_cleanup();
-void finish_indexer(char *script, char *index_id);
+void finish_indexer(char *script, int async_script, char *index_id);
 
 void elastic_init(int force_reset);
 
@@ -31,6 +31,6 @@ cJSON *elastic_get_document(const char *uuid_str);
 
 char *elastic_get_status();
 
-void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]);
+void execute_update_script(const char *script, int async, const char index_id[UUID_STR_LEN]);
 
 #endif

src/index/web.c

@@ -1,5 +1,6 @@
 #include "web.h"
 #include "src/sist.h"
+#include "src/ctx.h"
 
 #include <mongoose.h>
 #include <pthread.h>
@@ -21,95 +22,82 @@ void free_response(response_t *resp) {
     free(resp);
 }
 
-#define SIST2_HEADERS "User-Agent: sist2\r\nContent-Type: application/json\r\n"
+void web_post_async_poll(subreq_ctx_t* req) {
+    fd_set fdread;
+    fd_set fdwrite;
+    fd_set fdexcep;
+    int maxfd = -1;
 
+    FD_ZERO(&fdread);
+    FD_ZERO(&fdwrite);
+    FD_ZERO(&fdexcep);
 
-void http_req_ev(struct mg_connection *nc, int ev, void *ptr) {
+    CURLMcode mc = curl_multi_fdset(req->multi, &fdread, &fdwrite, &fdexcep, &maxfd);
 
-    http_ev_data_t *ev_data = (http_ev_data_t *) nc->user_data;
+    if(mc != CURLM_OK) {
+        req->done = TRUE;
+        return;
+    }
 
-    switch (ev) {
+    if (maxfd == -1) {
-        case MG_EV_CONNECT: {
+        // no fds ready yet
-            int connect_status = *(int *) ptr;
+        return;
-            if (connect_status != 0) {
+    }
-                ev_data->done = TRUE;
+
-                ev_data->resp->status_code = 0;
+    struct timeval timeout = {1, 0};
-            }
+    int rc = select(maxfd + 1, &fdread, &fdwrite, &fdexcep, &timeout);
+
+    switch(rc) {
+        case -1:
+            req->done = TRUE;
             break;
-        }
+        case 0:
-        case MG_EV_HTTP_REPLY: {
-            struct http_message *hm = (struct http_message *) ptr;
-
-            //TODO: Check errors?
-
-            ev_data->resp->size = hm->body.len;
-            ev_data->resp->status_code = hm->resp_code;
-            ev_data->resp->body = malloc(hm->body.len + 1);
-            memcpy(ev_data->resp->body, hm->body.p, hm->body.len);
-            *(ev_data->resp->body + hm->body.len) = '\0';
-
-            ev_data->done = TRUE;
             break;
-        }
-        case MG_EV_CLOSE: {
-            ev_data->done = TRUE;
-            break;
-        }
         default:
+            curl_multi_perform(req->multi, &req->running_handles);
             break;
     }
+
+    if (req->running_handles == 0) {
+        req->done = TRUE;
+        req->response->body = req->response_buf.buf;
+        req->response->size = req->response_buf.cur;
+        curl_easy_getinfo(req->handle, CURLINFO_RESPONSE_CODE, &req->response->status_code);
+
+        curl_multi_cleanup(req->multi);
+        curl_easy_cleanup(req->handle);
+        curl_slist_free_all(req->headers);
+        return;
+    }
 }
 
-subreq_ctx_t *http_req(const char *url, const char *extra_headers, const char *post_data, const char *method) {
+subreq_ctx_t *web_post_async(const char *url, char *data) {
+    subreq_ctx_t *req = calloc(1, sizeof(subreq_ctx_t));
+    req->response = calloc(1, sizeof(response_t));
+    req->data = data;
+    req->response_buf = dyn_buffer_create();
 
-    struct mg_str scheme;
+    req->handle = curl_easy_init();
-    struct mg_str user_info;
+    CURL *curl = req->handle;
-    struct mg_str host;
+    curl_easy_setopt(curl, CURLOPT_URL, url);
-    unsigned int port;
+    curl_easy_setopt(curl, CURLOPT_WRITEDATA, (void *) (&req->response_buf));
-    struct mg_str path;
+    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_cb);
-    struct mg_str query;
+    curl_easy_setopt(curl, CURLOPT_POST, 1);
-    struct mg_str fragment;
+    curl_easy_setopt(curl, CURLOPT_USERAGENT, "sist2");
 
-    if (post_data == NULL) post_data = "";
+    struct curl_slist *headers = NULL;
-    if (extra_headers == NULL) extra_headers = "";
+    headers = curl_slist_append(headers, "Content-Type: application/json");
-    if (path.len == 0) path = mg_mk_str("/");
+    curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
-    if (host.len == 0) host = mg_mk_str("");
 
-    // [scheme://[user_info@]]host[:port][/path][?query][#fragment]
+    curl_easy_setopt(curl, CURLOPT_POSTFIELDS, data);
-    mg_parse_uri(mg_mk_str(url), &scheme, &user_info, &host, &port, &path, &query, &fragment);
 
-    if (query.len > 0) path.len += query.len + 1;
+    req->multi = curl_multi_init();
+    curl_multi_add_handle(req->multi, curl);
+    curl_multi_perform(req->multi, &req->running_handles);
 
-    subreq_ctx_t *ctx = malloc(sizeof(subreq_ctx_t));
+    LOG_DEBUGF("web.c", "async request POST %s", url)
-    mg_mgr_init(&ctx->mgr, NULL);
 
-    char address[8192];
+    return req;
-    snprintf(address, sizeof(address), "tcp://%.*s:%u", (int) host.len, host.p, port);
-    struct mg_connection *nc = mg_connect(&ctx->mgr, address, http_req_ev);
-    nc->user_data = &ctx->ev_data;
-    mg_set_protocol_http_websocket(nc);
-
-    ctx->ev_data.resp = calloc(1, sizeof(response_t));
-    ctx->ev_data.done = FALSE;
-
-    mg_printf(
-            nc, "%s %.*s HTTP/1.1\r\n"
-                  "Host: %.*s\r\n"
-                  "Content-Length: %zu\r\n"
-                  "%s\r\n"
-                  "%s",
-            method, (int) path.len, path.p,
-            (int) (path.p - host.p), host.p,
-            strlen(post_data),
-            extra_headers,
-            post_data
-    );
-
-    return ctx;
-}
-
-subreq_ctx_t *web_post_async(const char *url, const char *data) {
-    return http_req(url, SIST2_HEADERS, data, "POST");
 }
 
 response_t *web_get(const char *url, int timeout) {
@@ -125,7 +113,8 @@ response_t *web_get(const char *url, int timeout) {
     curl_easy_setopt(curl, CURLOPT_USERAGENT, "sist2");
     curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout);
 
-    struct curl_slist *headers = curl_slist_append(headers, "Content-Type: application/json");
+    struct curl_slist *headers = NULL;
+    headers = curl_slist_append(headers, "Content-Type: application/json");
     curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
 
     curl_easy_perform(curl);
@@ -153,7 +142,8 @@ response_t *web_post(const char *url, const char *data) {
     curl_easy_setopt(curl, CURLOPT_POST, 1);
     curl_easy_setopt(curl, CURLOPT_USERAGENT, "sist2");
 
-    struct curl_slist *headers = curl_slist_append(headers, "Content-Type: application/json");
+    struct curl_slist *headers = NULL;
+    headers = curl_slist_append(headers, "Content-Type: application/json");
     curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
 
     curl_easy_setopt(curl, CURLOPT_POSTFIELDS, data);
@@ -187,7 +177,8 @@ response_t *web_put(const char *url, const char *data) {
     curl_easy_setopt(curl, CURLOPT_DNS_USE_GLOBAL_CACHE, 0);
     curl_easy_setopt(curl, CURLOPT_IPRESOLVE, CURLOPT_DNS_LOCAL_IP4 );
 
-    struct curl_slist *headers = curl_slist_append(headers, "Content-Type: application/json");
+    struct curl_slist *headers = NULL;
+    headers = curl_slist_append(headers, "Content-Type: application/json");
     curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
 
     curl_easy_setopt(curl, CURLOPT_POSTFIELDS, data);
@@ -218,7 +209,8 @@ response_t *web_delete(const char *url) {
     curl_easy_setopt(curl, CURLOPT_USERAGENT, "sist2");
 
     curl_easy_setopt(curl, CURLOPT_POSTFIELDS, "");
-    struct curl_slist *headers = curl_slist_append(headers, "Content-Type: application/json");
+    struct curl_slist *headers = NULL;
+    headers = curl_slist_append(headers, "Content-Type: application/json");
     curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
 
     curl_easy_perform(curl);

src/index/web.h

@@ -3,6 +3,7 @@
 
 #include "src/sist.h"
 #include <mongoose.h>
+#include <curl/curl.h>
 
 typedef struct response {
     char *body;
@@ -16,13 +17,20 @@ typedef struct {
 } http_ev_data_t;
 
 typedef struct {
-    http_ev_data_t ev_data;
+    char* data;
-    struct mg_mgr mgr;
+    dyn_buffer_t response_buf;
+    struct curl_slist *headers;
+    CURL *handle;
+    CURLM *multi;
+    response_t *response;
+    int running_handles;
+    int done;
 } subreq_ctx_t;
 
 response_t *web_get(const char *url, int timeout);
 response_t *web_post(const char * url, const char * data);
-subreq_ctx_t *web_post_async(const char *url, const char *data);
+void web_post_async_poll(subreq_ctx_t* req);
+subreq_ctx_t *web_post_async(const char *url, char *data);
 response_t *web_put(const char *url, const char *data);
 response_t *web_delete(const char *url);
 

src/io/serialize.c

@@ -150,6 +150,8 @@ char *get_meta_key_text(enum metakey meta_key) {
             return "modified_by";
         case MetaThumbnail:
             return "thumbnail";
+        case MetaPages:
+            return "pages";
         default:
             return NULL;
     }
@@ -278,6 +280,7 @@ void read_index_bin(const char *path, const char *index_id, index_func func) {
         size_t ret = 0;
         while (key != '\n') {
             switch (key) {
+                case MetaPages:
                 case MetaWidth:
                 case MetaHeight: {
                     int value;

src/log.c

@@ -4,8 +4,7 @@
 #include <stdarg.h>
 
 const char *log_colors[] = {
-        "\033[34m", "\033[01;34m", "\033[0m",
+        "\033[34m", "\033[01;34m", "\033[01;33m", "\033[0m", "\033[31m", "\033[01;31m"
-        "\033[01;33m", "\033[31m", "\033[01;31m"
 };
 
 const char *log_levels[] = {

src/main.c

@@ -21,7 +21,7 @@
 #define EPILOG "Made by simon987 <me@simon987.net>. Released under GPL-3.0"
 
 
-static const char *const Version = "2.7.4";
+static const char *const Version = "2.8.4";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",
@@ -259,6 +259,7 @@ void sist2_scan(scan_args_t *args) {
 void sist2_index(index_args_t *args) {
 
     IndexCtx.es_url = args->es_url;
+    IndexCtx.es_index = args->es_index;
     IndexCtx.batch_size = args->batch_size;
 
     if (!args->print) {
@@ -317,11 +318,12 @@ void sist2_index(index_args_t *args) {
 
     tpool_wait(IndexCtx.pool);
 
-    if (!args->print) {
-        finish_indexer(args->script, desc.uuid);
-    }
     tpool_destroy(IndexCtx.pool);
 
+    if (!args->print) {
+        finish_indexer(args->script, args->async_script, desc.uuid);
+    }
+
     store_destroy(IndexCtx.tag_store);
     g_hash_table_remove_all(IndexCtx.tags);
     g_hash_table_destroy(IndexCtx.tags);
@@ -339,13 +341,14 @@ void sist2_exec_script(exec_args_t *args) {
 
     LOG_DEBUGF("main.c", "descriptor version %s (%s)", desc.version, desc.type)
 
-    execute_update_script(args->script, desc.uuid);
+    execute_update_script(args->script, args->async_script, desc.uuid);
     free(args->script);
 }
 
 void sist2_web(web_args_t *args) {
 
     WebCtx.es_url = args->es_url;
+    WebCtx.es_index = args->es_index;
     WebCtx.index_count = args->index_count;
     WebCtx.auth_user = args->auth_user;
     WebCtx.auth_pass = args->auth_pass;
@@ -389,7 +392,9 @@ int main(int argc, const char *argv[]) {
     int arg_version = 0;
 
     char *common_es_url = NULL;
+    char *common_es_index = NULL;
     char *common_script_path = NULL;
+    int common_async_script = 0;
     int common_threads = 0;
 
     struct argparse_option options[] = {
@@ -430,20 +435,26 @@ int main(int argc, const char *argv[]) {
             OPT_GROUP("Index options"),
             OPT_INTEGER('t', "threads", &common_threads, "Number of threads. DEFAULT=1"),
             OPT_STRING(0, "es-url", &common_es_url, "Elasticsearch url with port. DEFAULT=http://localhost:9200"),
+            OPT_STRING(0, "es-index", &common_es_index, "Elasticsearch index name. DEFAULT=sist2"),
             OPT_BOOLEAN('p', "print", &index_args->print, "Just print JSON documents to stdout."),
             OPT_STRING(0, "script-file", &common_script_path, "Path to user script."),
+            OPT_BOOLEAN(0, "async-script", &common_async_script, "Execute user script asynchronously."),
             OPT_INTEGER(0, "batch-size", &index_args->batch_size, "Index batch size. DEFAULT: 100"),
             OPT_BOOLEAN('f', "force-reset", &index_args->force_reset, "Reset Elasticsearch mappings and settings. "
                                                                       "(You must use this option the first time you use the index command)"),
 
             OPT_GROUP("Web options"),
             OPT_STRING(0, "es-url", &common_es_url, "Elasticsearch url. DEFAULT=http://localhost:9200"),
+            OPT_STRING(0, "es-index", &common_es_index, "Elasticsearch index name. DEFAULT=sist2"),
             OPT_STRING(0, "bind", &web_args->listen_address, "Listen on this address. DEFAULT=localhost:4090"),
             OPT_STRING(0, "auth", &web_args->credentials, "Basic auth in user:password format"),
             OPT_STRING(0, "tag-auth", &web_args->tag_credentials, "Basic auth in user:password format for tagging"),
 
             OPT_GROUP("Exec-script options"),
+            OPT_STRING(0, "es-url", &common_es_url, "Elasticsearch url. DEFAULT=http://localhost:9200"),
+            OPT_STRING(0, "es-index", &common_es_index, "Elasticsearch index name. DEFAULT=sist2"),
             OPT_STRING(0, "script-file", &common_script_path, "Path to user script."),
+            OPT_BOOLEAN(0, "async-script", &common_async_script, "Execute user script asynchronously."),
 
             OPT_END(),
     };
@@ -465,10 +476,17 @@ int main(int argc, const char *argv[]) {
     web_args->es_url = common_es_url;
     index_args->es_url = common_es_url;
     exec_args->es_url = common_es_url;
+
+    web_args->es_index = common_es_index;
+    index_args->es_index = common_es_index;
+    exec_args->es_index = common_es_index;
+
     index_args->script_path = common_script_path;
     exec_args->script_path = common_script_path;
     index_args->threads = common_threads;
     scan_args->threads = common_threads;
+    exec_args->async_script = common_async_script;
+    index_args->async_script = common_async_script;
 
     if (argc == 0) {
         argparse_usage(&argparse);

src/parsing/mime_generated.c

@@ -54,406 +54,407 @@ enum mime {
     application_streamingmedia=655406,
     application_vda=655407,
     application_vnd_amazon_mobi8_ebook=655408 | 0x02000000,
-    application_vnd_fdf=655409,
+    application_vnd_coffeescript=655409,
-    application_vnd_font_fontforge_sfd=655410,
+    application_vnd_fdf=655410,
-    application_vnd_hp_hpgl=655411,
+    application_vnd_font_fontforge_sfd=655411,
-    application_vnd_iccprofile=655412,
+    application_vnd_hp_hpgl=655412,
-    application_vnd_lotus_1_2_3=655413,
+    application_vnd_iccprofile=655413,
-    application_vnd_ms_cab_compressed=655414,
+    application_vnd_lotus_1_2_3=655414,
-    application_vnd_ms_excel=655415,
+    application_vnd_ms_cab_compressed=655415,
-    application_vnd_ms_fontobject=655416,
+    application_vnd_ms_excel=655416,
-    application_vnd_ms_opentype=655417 | 0x20000000,
+    application_vnd_ms_fontobject=655417,
-    application_vnd_ms_outlook=655418,
+    application_vnd_ms_opentype=655418 | 0x20000000,
-    application_vnd_ms_pki_certstore=655419,
+    application_vnd_ms_outlook=655419,
-    application_vnd_ms_pki_pko=655420,
+    application_vnd_ms_pki_certstore=655420,
-    application_vnd_ms_pki_seccat=655421,
+    application_vnd_ms_pki_pko=655421,
-    application_vnd_ms_powerpoint=655422,
+    application_vnd_ms_pki_seccat=655422,
-    application_vnd_ms_project=655423,
+    application_vnd_ms_powerpoint=655423,
-    application_vnd_oasis_opendocument_base=655424,
+    application_vnd_ms_project=655424,
-    application_vnd_oasis_opendocument_formula=655425,
+    application_vnd_oasis_opendocument_base=655425,
-    application_vnd_oasis_opendocument_graphics=655426,
+    application_vnd_oasis_opendocument_formula=655426,
-    application_vnd_oasis_opendocument_presentation=655427,
+    application_vnd_oasis_opendocument_graphics=655427,
-    application_vnd_oasis_opendocument_spreadsheet=655428,
+    application_vnd_oasis_opendocument_presentation=655428,
-    application_vnd_oasis_opendocument_text=655429,
+    application_vnd_oasis_opendocument_spreadsheet=655429,
-    application_vnd_openxmlformats_officedocument_presentationml_presentation=655430 | 0x04000000,
+    application_vnd_oasis_opendocument_text=655430,
-    application_vnd_openxmlformats_officedocument_spreadsheetml_sheet=655431 | 0x04000000,
+    application_vnd_openxmlformats_officedocument_presentationml_presentation=655431 | 0x04000000,
-    application_vnd_openxmlformats_officedocument_wordprocessingml_document=655432 | 0x04000000,
+    application_vnd_openxmlformats_officedocument_spreadsheetml_sheet=655432 | 0x04000000,
-    application_vnd_symbian_install=655433,
+    application_vnd_openxmlformats_officedocument_wordprocessingml_document=655433 | 0x04000000,
-    application_vnd_tcpdump_pcap=655434,
+    application_vnd_symbian_install=655434,
-    application_vnd_wap_wmlc=655435,
+    application_vnd_tcpdump_pcap=655435,
-    application_vnd_wap_wmlscriptc=655436,
+    application_vnd_wap_wmlc=655436,
-    application_vnd_xara=655437,
+    application_vnd_wap_wmlscriptc=655437,
-    application_vocaltec_media_desc=655438,
+    application_vnd_xara=655438,
-    application_vocaltec_media_file=655439,
+    application_vocaltec_media_desc=655439,
-    application_warc=655440,
+    application_vocaltec_media_file=655440,
-    application_winhelp=655441,
+    application_warc=655441,
-    application_wordperfect=655442,
+    application_winhelp=655442,
-    application_wordperfect6_0=655443,
+    application_wordperfect=655443,
-    application_wordperfect6_1=655444,
+    application_wordperfect6_0=655444,
-    application_x_123=655445,
+    application_wordperfect6_1=655445,
-    application_x_7z_compressed=655446 | 0x10000000,
+    application_x_123=655446,
-    application_x_aim=655447,
+    application_x_7z_compressed=655447 | 0x10000000,
-    application_x_apple_diskimage=655448,
+    application_x_aim=655448,
-    application_x_arc=655449 | 0x10000000,
+    application_x_apple_diskimage=655449,
-    application_x_archive=655450,
+    application_x_arc=655450 | 0x10000000,
-    application_x_atari_7800_rom=655451,
+    application_x_archive=655451,
-    application_x_authorware_bin=655452,
+    application_x_atari_7800_rom=655452,
-    application_x_authorware_map=655453,
+    application_x_authorware_bin=655453,
-    application_x_authorware_seg=655454,
+    application_x_authorware_map=655454,
-    application_x_avira_qua=655455,
+    application_x_authorware_seg=655455,
-    application_x_bcpio=655456,
+    application_x_avira_qua=655456,
-    application_x_bittorrent=655457,
+    application_x_bcpio=655457,
-    application_x_bsh=655458,
+    application_x_bittorrent=655458,
-    application_x_bytecode_python=655459,
+    application_x_bsh=655459,
-    application_x_bzip=655460,
+    application_x_bytecode_python=655460,
-    application_x_bzip2=655461 | 0x08000000,
+    application_x_bzip=655461,
-    application_x_cbr=655462,
+    application_x_bzip2=655462 | 0x08000000,
-    application_x_cbz=655463,
+    application_x_cbr=655463,
-    application_x_cdlink=655464,
+    application_x_cbz=655464,
-    application_x_chat=655465,
+    application_x_cdlink=655465,
-    application_x_chrome_extension=655466,
+    application_x_chat=655466,
-    application_x_cocoa=655467,
+    application_x_chrome_extension=655467,
-    application_x_conference=655468,
+    application_x_cocoa=655468,
-    application_x_coredump=655469,
+    application_x_conference=655469,
-    application_x_cpio=655470,
+    application_x_coredump=655470,
-    application_x_dbf=655471,
+    application_x_cpio=655471,
-    application_x_dbt=655472,
+    application_x_dbf=655472,
-    application_x_debian_package=655473,
+    application_x_dbt=655473,
-    application_x_deepv=655474,
+    application_x_debian_package=655474,
-    application_x_director=655475,
+    application_x_deepv=655475,
-    application_x_dmp=655476,
+    application_x_director=655476,
-    application_x_dosdriver=655477,
+    application_x_dmp=655477,
-    application_x_dosexec=655478,
+    application_x_dosdriver=655478,
-    application_x_dvi=655479,
+    application_x_dosexec=655479,
-    application_x_elc=655480,
+    application_x_dvi=655480,
+    application_x_elc=655481,
     application_x_empty=1,
-    application_x_envoy=655482,
+    application_x_envoy=655483,
-    application_x_esrehber=655483,
+    application_x_esrehber=655484,
-    application_x_excel=655484,
+    application_x_excel=655485,
-    application_x_executable=655485,
+    application_x_executable=655486,
-    application_x_font_gdos=655486,
+    application_x_font_gdos=655487,
-    application_x_font_pf2=655487,
+    application_x_font_pf2=655488,
-    application_x_font_pfm=655488,
+    application_x_font_pfm=655489,
-    application_x_font_sfn=655489,
+    application_x_font_sfn=655490,
-    application_x_font_ttf=655490 | 0x20000000,
+    application_x_font_ttf=655491 | 0x20000000,
-    application_x_fptapplication_x_dbt=655491,
+    application_x_fptapplication_x_dbt=655492,
-    application_x_freelance=655492,
+    application_x_freelance=655493,
-    application_x_gamecube_rom=655493,
+    application_x_gamecube_rom=655494,
-    application_x_gdbm=655494,
+    application_x_gdbm=655495,
-    application_x_gettext_translation=655495,
+    application_x_gettext_translation=655496,
-    application_x_git=655496,
+    application_x_git=655497,
-    application_x_gsp=655497,
+    application_x_gsp=655498,
-    application_x_gss=655498,
+    application_x_gss=655499,
-    application_x_gtar=655499,
+    application_x_gtar=655500,
-    application_x_gzip=655500,
+    application_x_gzip=655501,
-    application_x_hdf=655501,
+    application_x_hdf=655502,
-    application_x_helpfile=655502,
+    application_x_helpfile=655503,
-    application_x_httpd_imap=655503,
+    application_x_httpd_imap=655504,
-    application_x_ima=655504,
+    application_x_ima=655505,
-    application_x_innosetup=655505,
+    application_x_innosetup=655506,
-    application_x_internett_signup=655506,
+    application_x_internett_signup=655507,
-    application_x_inventor=655507,
+    application_x_inventor=655508,
-    application_x_ip2=655508,
+    application_x_ip2=655509,
-    application_x_java_applet=655509,
+    application_x_java_applet=655510,
-    application_x_java_commerce=655510,
+    application_x_java_commerce=655511,
-    application_x_java_image=655511,
+    application_x_java_image=655512,
-    application_x_java_jmod=655512,
+    application_x_java_jmod=655513,
-    application_x_java_keystore=655513,
+    application_x_java_keystore=655514,
-    application_x_kdelnk=655514,
+    application_x_kdelnk=655515,
-    application_x_koan=655515,
+    application_x_koan=655516,
-    application_x_latex=655516,
+    application_x_latex=655517,
-    application_x_livescreen=655517,
+    application_x_livescreen=655518,
-    application_x_lotus=655518,
+    application_x_lotus=655519,
-    application_x_lz4=655519 | 0x08000000,
+    application_x_lz4=655520 | 0x08000000,
-    application_x_lz4_json=655520,
+    application_x_lz4_json=655521,
-    application_x_lzh=655521,
+    application_x_lzh=655522,
-    application_x_lzh_compressed=655522,
+    application_x_lzh_compressed=655523,
-    application_x_lzip=655523 | 0x08000000,
+    application_x_lzip=655524 | 0x08000000,
-    application_x_lzma=655524 | 0x08000000,
+    application_x_lzma=655525 | 0x08000000,
-    application_x_lzop=655525 | 0x08000000,
+    application_x_lzop=655526 | 0x08000000,
-    application_x_lzx=655526,
+    application_x_lzx=655527,
-    application_x_mach_binary=655527,
+    application_x_mach_binary=655528,
-    application_x_mach_executable=655528,
+    application_x_mach_executable=655529,
-    application_x_magic_cap_package_1_0=655529,
+    application_x_magic_cap_package_1_0=655530,
-    application_x_mathcad=655530,
+    application_x_mathcad=655531,
-    application_x_maxis_dbpf=655531,
+    application_x_maxis_dbpf=655532,
-    application_x_meme=655532,
+    application_x_meme=655533,
-    application_x_midi=655533,
+    application_x_midi=655534,
-    application_x_mif=655534,
+    application_x_mif=655535,
-    application_x_mix_transfer=655535,
+    application_x_mix_transfer=655536,
-    application_x_mobipocket_ebook=655536 | 0x02000000,
+    application_x_mobipocket_ebook=655537 | 0x02000000,
-    application_x_ms_compress_szdd=655537,
+    application_x_ms_compress_szdd=655538,
-    application_x_ms_pdb=655538,
+    application_x_ms_pdb=655539,
-    application_x_ms_reader=655539,
+    application_x_ms_reader=655540,
-    application_x_msaccess=655540,
+    application_x_msaccess=655541,
-    application_x_n64_rom=655541,
+    application_x_n64_rom=655542,
-    application_x_navi_animation=655542,
+    application_x_navi_animation=655543,
-    application_x_navidoc=655543,
+    application_x_navidoc=655544,
-    application_x_navimap=655544,
+    application_x_navimap=655545,
-    application_x_navistyle=655545,
+    application_x_navistyle=655546,
-    application_x_nes_rom=655546,
+    application_x_nes_rom=655547,
-    application_x_netcdf=655547,
+    application_x_netcdf=655548,
-    application_x_newton_compatible_pkg=655548,
+    application_x_newton_compatible_pkg=655549,
-    application_x_nintendo_ds_rom=655549,
+    application_x_nintendo_ds_rom=655550,
-    application_x_object=655550,
+    application_x_object=655551,
-    application_x_omc=655551,
+    application_x_omc=655552,
-    application_x_omcdatamaker=655552,
+    application_x_omcdatamaker=655553,
-    application_x_omcregerator=655553,
+    application_x_omcregerator=655554,
-    application_x_pagemaker=655554,
+    application_x_pagemaker=655555,
-    application_x_pcl=655555,
+    application_x_pcl=655556,
-    application_x_pgp_keyring=655556,
+    application_x_pgp_keyring=655557,
-    application_x_pixclscript=655557,
+    application_x_pixclscript=655558,
-    application_x_pkcs7_certreqresp=655558,
+    application_x_pkcs7_certreqresp=655559,
-    application_x_pkcs7_signature=655559,
+    application_x_pkcs7_signature=655560,
-    application_x_project=655560,
+    application_x_project=655561,
-    application_x_qpro=655561,
+    application_x_qpro=655562,
-    application_x_rar=655562 | 0x10000000,
+    application_x_rar=655563 | 0x10000000,
-    application_x_rpm=655563,
+    application_x_rpm=655564,
-    application_x_sdp=655564,
+    application_x_sdp=655565,
-    application_x_sea=655565,
+    application_x_sea=655566,
-    application_x_seelogo=655566,
+    application_x_seelogo=655567,
-    application_x_setupscript=655567,
+    application_x_setupscript=655568,
-    application_x_shar=655568,
+    application_x_shar=655569,
-    application_x_sharedlib=655569,
+    application_x_sharedlib=655570,
-    application_x_shockwave_flash=655570,
+    application_x_shockwave_flash=655571,
-    application_x_snappy_framed=655571,
+    application_x_snappy_framed=655572,
-    application_x_sprite=655572,
+    application_x_sprite=655573,
-    application_x_sqlite3=655573,
+    application_x_sqlite3=655574,
-    application_x_stargallery_thm=655574,
+    application_x_stargallery_thm=655575,
-    application_x_stuffit=655575,
+    application_x_stuffit=655576,
-    application_x_sv4cpio=655576,
+    application_x_sv4cpio=655577,
-    application_x_sv4crc=655577,
+    application_x_sv4crc=655578,
-    application_x_tar=655578 | 0x10000000,
+    application_x_tar=655579 | 0x10000000,
-    application_x_tbook=655579,
+    application_x_tbook=655580,
-    application_x_terminfo=655580,
+    application_x_terminfo=655581,
-    application_x_terminfo2=655581,
+    application_x_terminfo2=655582,
-    application_x_tex_tfm=655582,
+    application_x_tex_tfm=655583,
-    application_x_texinfo=655583,
+    application_x_texinfo=655584,
-    application_x_ustar=655584,
+    application_x_ustar=655585,
-    application_x_visio=655585,
+    application_x_visio=655586,
-    application_x_vnd_audioexplosion_mzz=655586,
+    application_x_vnd_audioexplosion_mzz=655587,
-    application_x_vnd_ls_xpix=655587,
+    application_x_vnd_ls_xpix=655588,
-    application_x_vrml=655588,
+    application_x_vrml=655589,
-    application_x_wais_source=655589,
+    application_x_wais_source=655590,
-    application_x_wine_extension_ini=655590,
+    application_x_wine_extension_ini=655591,
-    application_x_wintalk=655591,
+    application_x_wintalk=655592,
-    application_x_world=655592,
+    application_x_world=655593,
-    application_x_wri=655593,
+    application_x_wri=655594,
-    application_x_x509_ca_cert=655594,
+    application_x_x509_ca_cert=655595,
-    application_x_xz=655595 | 0x08000000,
+    application_x_xz=655596 | 0x08000000,
-    application_x_zip=655596,
+    application_x_zip=655597,
-    application_x_zstd=655597 | 0x08000000,
+    application_x_zstd=655598 | 0x08000000,
-    application_x_zstd_dictionary=655598,
+    application_x_zstd_dictionary=655599,
-    application_xml=655599,
+    application_xml=655600,
-    application_zip=655600 | 0x10000000,
+    application_zip=655601 | 0x10000000,
-    application_zlib=655601,
+    application_zlib=655602,
-    audio_basic=458994 | 0x80000000,
+    audio_basic=458995 | 0x80000000,
-    audio_it=458995,
+    audio_it=458996,
-    audio_make=458996,
+    audio_make=458997,
-    audio_mid=458997,
+    audio_mid=458998,
-    audio_midi=458998,
+    audio_midi=458999,
-    audio_mp4=458999,
+    audio_mp4=459000,
-    audio_mpeg=459000,
+    audio_mpeg=459001,
-    audio_ogg=459001,
+    audio_ogg=459002,
-    audio_s3m=459002,
+    audio_s3m=459003,
-    audio_tsp_audio=459003,
+    audio_tsp_audio=459004,
-    audio_tsplayer=459004,
+    audio_tsplayer=459005,
-    audio_vnd_qcelp=459005,
+    audio_vnd_qcelp=459006,
-    audio_voxware=459006,
+    audio_voxware=459007,
-    audio_x_aiff=459007,
+    audio_x_aiff=459008,
-    audio_x_flac=459008,
+    audio_x_flac=459009,
-    audio_x_gsm=459009,
+    audio_x_gsm=459010,
-    audio_x_hx_aac_adts=459010,
+    audio_x_hx_aac_adts=459011,
-    audio_x_jam=459011,
+    audio_x_jam=459012,
-    audio_x_liveaudio=459012,
+    audio_x_liveaudio=459013,
-    audio_x_m4a=459013,
+    audio_x_m4a=459014,
-    audio_x_midi=459014,
+    audio_x_midi=459015,
-    audio_x_mod=459015,
+    audio_x_mod=459016,
-    audio_x_mp4a_latm=459016,
+    audio_x_mp4a_latm=459017,
-    audio_x_mpeg_3=459017,
+    audio_x_mpeg_3=459018,
-    audio_x_mpequrl=459018,
+    audio_x_mpequrl=459019,
-    audio_x_nspaudio=459019,
+    audio_x_nspaudio=459020,
-    audio_x_pn_realaudio=459020,
+    audio_x_pn_realaudio=459021,
-    audio_x_psid=459021,
+    audio_x_psid=459022,
-    audio_x_realaudio=459022,
+    audio_x_realaudio=459023,
-    audio_x_s3m=459023,
+    audio_x_s3m=459024,
-    audio_x_twinvq=459024,
+    audio_x_twinvq=459025,
-    audio_x_twinvq_plugin=459025,
+    audio_x_twinvq_plugin=459026,
-    audio_x_voc=459026,
+    audio_x_voc=459027,
-    audio_x_wav=459027,
+    audio_x_wav=459028,
-    audio_x_xbox_executable=459028 | 0x80000000,
+    audio_x_xbox_executable=459029 | 0x80000000,
-    audio_x_xbox360_executable=459029 | 0x80000000,
+    audio_x_xbox360_executable=459030 | 0x80000000,
-    audio_xm=459030,
+    audio_xm=459031,
-    font_otf=327959 | 0x20000000,
+    font_otf=327960 | 0x20000000,
-    font_sfnt=327960 | 0x20000000,
+    font_sfnt=327961 | 0x20000000,
-    font_woff=327961 | 0x20000000,
+    font_woff=327962 | 0x20000000,
-    font_woff2=327962 | 0x20000000,
+    font_woff2=327963 | 0x20000000,
-    image_bmp=524571,
+    image_bmp=524572,
-    image_cmu_raster=524572,
+    image_cmu_raster=524573,
-    image_fif=524573,
+    image_fif=524574,
-    image_florian=524574,
+    image_florian=524575,
-    image_g3fax=524575,
+    image_g3fax=524576,
-    image_gif=524576,
+    image_gif=524577,
-    image_heic=524577,
+    image_heic=524578,
-    image_ief=524578,
+    image_ief=524579,
-    image_jpeg=524579,
+    image_jpeg=524580,
-    image_jutvision=524580,
+    image_jutvision=524581,
-    image_naplps=524581,
+    image_naplps=524582,
-    image_pict=524582,
+    image_pict=524583,
-    image_png=524583,
+    image_png=524584,
-    image_svg=524584 | 0x80000000,
+    image_svg=524585 | 0x80000000,
-    image_svg_xml=524585 | 0x80000000,
+    image_svg_xml=524586 | 0x80000000,
-    image_tiff=524586,
+    image_tiff=524587,
-    image_vnd_adobe_photoshop=524587 | 0x80000000,
+    image_vnd_adobe_photoshop=524588 | 0x80000000,
-    image_vnd_djvu=524588 | 0x80000000,
+    image_vnd_djvu=524589 | 0x80000000,
-    image_vnd_fpx=524589,
+    image_vnd_fpx=524590,
-    image_vnd_microsoft_icon=524590,
+    image_vnd_microsoft_icon=524591,
-    image_vnd_rn_realflash=524591,
+    image_vnd_rn_realflash=524592,
-    image_vnd_rn_realpix=524592,
+    image_vnd_rn_realpix=524593,
-    image_vnd_wap_wbmp=524593,
+    image_vnd_wap_wbmp=524594,
-    image_vnd_xiff=524594,
+    image_vnd_xiff=524595,
-    image_webp=524595,
+    image_webp=524596,
-    image_wmf=524596,
+    image_wmf=524597,
-    image_x_3ds=524597,
+    image_x_3ds=524598,
-    image_x_adobe_dng=524598 | 0x00800000,
+    image_x_adobe_dng=524599 | 0x00800000,
-    image_x_award_bioslogo=524599,
+    image_x_award_bioslogo=524600,
-    image_x_canon_cr2=524600 | 0x00800000,
+    image_x_canon_cr2=524601 | 0x00800000,
-    image_x_canon_crw=524601 | 0x00800000,
+    image_x_canon_crw=524602 | 0x00800000,
-    image_x_cmu_raster=524602,
+    image_x_cmu_raster=524603,
-    image_x_cur=524603,
+    image_x_cur=524604,
-    image_x_dcraw=524604 | 0x00800000,
+    image_x_dcraw=524605 | 0x00800000,
-    image_x_dwg=524605,
+    image_x_dwg=524606,
-    image_x_eps=524606,
+    image_x_eps=524607,
-    image_x_epson_erf=524607 | 0x00800000,
+    image_x_epson_erf=524608 | 0x00800000,
-    image_x_exr=524608,
+    image_x_exr=524609,
-    image_x_fuji_raf=524609 | 0x00800000,
+    image_x_fuji_raf=524610 | 0x00800000,
-    image_x_gem=524610,
+    image_x_gem=524611,
-    image_x_icns=524611,
+    image_x_icns=524612,
-    image_x_icon=524612 | 0x80000000,
+    image_x_icon=524613 | 0x80000000,
-    image_x_jg=524613,
+    image_x_jg=524614,
-    image_x_jps=524614,
+    image_x_jps=524615,
-    image_x_kodak_dcr=524615 | 0x00800000,
+    image_x_kodak_dcr=524616 | 0x00800000,
-    image_x_kodak_k25=524616 | 0x00800000,
+    image_x_kodak_k25=524617 | 0x00800000,
-    image_x_kodak_kdc=524617 | 0x00800000,
+    image_x_kodak_kdc=524618 | 0x00800000,
-    image_x_minolta_mrw=524618 | 0x00800000,
+    image_x_minolta_mrw=524619 | 0x00800000,
-    image_x_ms_bmp=524619,
+    image_x_ms_bmp=524620,
-    image_x_niff=524620,
+    image_x_niff=524621,
-    image_x_nikon_nef=524621 | 0x00800000,
+    image_x_nikon_nef=524622 | 0x00800000,
-    image_x_olympus_orf=524622 | 0x00800000,
+    image_x_olympus_orf=524623 | 0x00800000,
-    image_x_panasonic_raw=524623 | 0x00800000,
+    image_x_panasonic_raw=524624 | 0x00800000,
-    image_x_pcx=524624,
+    image_x_pcx=524625,
-    image_x_pentax_pef=524625 | 0x00800000,
+    image_x_pentax_pef=524626 | 0x00800000,
-    image_x_pict=524626,
+    image_x_pict=524627,
-    image_x_portable_bitmap=524627,
+    image_x_portable_bitmap=524628,
-    image_x_portable_graymap=524628,
+    image_x_portable_graymap=524629,
-    image_x_portable_pixmap=524629,
+    image_x_portable_pixmap=524630,
-    image_x_quicktime=524630,
+    image_x_quicktime=524631,
-    image_x_rgb=524631,
+    image_x_rgb=524632,
-    image_x_sigma_x3f=524632 | 0x00800000,
+    image_x_sigma_x3f=524633 | 0x00800000,
-    image_x_sony_arw=524633 | 0x00800000,
+    image_x_sony_arw=524634 | 0x00800000,
-    image_x_sony_sr2=524634 | 0x00800000,
+    image_x_sony_sr2=524635 | 0x00800000,
-    image_x_sony_srf=524635 | 0x00800000,
+    image_x_sony_srf=524636 | 0x00800000,
-    image_x_tga=524636,
+    image_x_tga=524637,
-    image_x_tiff=524637,
+    image_x_tiff=524638,
-    image_x_win_bitmap=524638,
+    image_x_win_bitmap=524639,
-    image_x_xcf=524639 | 0x80000000,
+    image_x_xcf=524640 | 0x80000000,
-    image_x_xpixmap=524640 | 0x80000000,
+    image_x_xpixmap=524641 | 0x80000000,
-    image_x_xwindowdump=524641,
+    image_x_xwindowdump=524642,
-    message_news=196962,
+    message_news=196963,
-    message_rfc822=196963,
+    message_rfc822=196964,
-    model_vnd_dwf=65892,
+    model_vnd_dwf=65893,
-    model_vnd_gdl=65893,
+    model_vnd_gdl=65894,
-    model_vnd_gs_gdl=65894,
+    model_vnd_gs_gdl=65895,
-    model_vrml=65895,
+    model_vrml=65896,
-    model_x_pov=65896,
+    model_x_pov=65897,
-    text_PGP=590185,
+    text_PGP=590186,
-    text_asp=590186,
+    text_asp=590187,
-    text_css=590187,
+    text_css=590188,
-    text_html=590188 | 0x01000000,
+    text_html=590189 | 0x01000000,
-    text_javascript=590189,
+    text_javascript=590190,
-    text_mcf=590190,
+    text_mcf=590191,
-    text_pascal=590191,
+    text_pascal=590192,
-    text_plain=590192,
+    text_plain=590193,
-    text_richtext=590193,
+    text_richtext=590194,
-    text_rtf=590194,
+    text_rtf=590195,
-    text_scriplet=590195,
+    text_scriplet=590196,
-    text_tab_separated_values=590196,
+    text_tab_separated_values=590197,
-    text_troff=590197,
+    text_troff=590198,
-    text_uri_list=590198,
+    text_uri_list=590199,
-    text_vnd_abc=590199,
+    text_vnd_abc=590200,
-    text_vnd_fmi_flexstor=590200,
+    text_vnd_fmi_flexstor=590201,
-    text_vnd_wap_wml=590201,
+    text_vnd_wap_wml=590202,
-    text_vnd_wap_wmlscript=590202,
+    text_vnd_wap_wmlscript=590203,
-    text_webviewhtml=590203,
+    text_webviewhtml=590204,
-    text_x_Algol68=590204,
+    text_x_Algol68=590205,
-    text_x_asm=590205,
+    text_x_asm=590206,
-    text_x_audiosoft_intra=590206,
+    text_x_audiosoft_intra=590207,
-    text_x_awk=590207,
+    text_x_awk=590208,
-    text_x_bcpl=590208,
+    text_x_bcpl=590209,
-    text_x_c=590209,
+    text_x_c=590210,
-    text_x_c__=590210,
+    text_x_c__=590211,
-    text_x_component=590211,
+    text_x_component=590212,
-    text_x_diff=590212,
+    text_x_diff=590213,
-    text_x_fortran=590213,
+    text_x_fortran=590214,
-    text_x_java=590214,
+    text_x_java=590215,
-    text_x_la_asf=590215,
+    text_x_la_asf=590216,
-    text_x_lisp=590216,
+    text_x_lisp=590217,
-    text_x_m=590217,
+    text_x_m=590218,
-    text_x_m4=590218,
+    text_x_m4=590219,
-    text_x_makefile=590219,
+    text_x_makefile=590220,
-    text_x_ms_regedit=590220,
+    text_x_ms_regedit=590221,
-    text_x_msdos_batch=590221,
+    text_x_msdos_batch=590222,
-    text_x_objective_c=590222,
+    text_x_objective_c=590223,
-    text_x_pascal=590223,
+    text_x_pascal=590224,
-    text_x_perl=590224,
+    text_x_perl=590225,
-    text_x_php=590225,
+    text_x_php=590226,
-    text_x_po=590226,
+    text_x_po=590227,
-    text_x_python=590227,
+    text_x_python=590228,
-    text_x_ruby=590228,
+    text_x_ruby=590229,
-    text_x_sass=590229,
+    text_x_sass=590230,
-    text_x_scss=590230,
+    text_x_scss=590231,
-    text_x_server_parsed_html=590231,
+    text_x_server_parsed_html=590232,
-    text_x_setext=590232,
+    text_x_setext=590233,
-    text_x_sgml=590233 | 0x01000000,
+    text_x_sgml=590234 | 0x01000000,
-    text_x_shellscript=590234,
+    text_x_shellscript=590235,
-    text_x_speech=590235,
+    text_x_speech=590236,
-    text_x_tcl=590236,
+    text_x_tcl=590237,
-    text_x_tex=590237,
+    text_x_tex=590238,
-    text_x_uil=590238,
+    text_x_uil=590239,
-    text_x_uuencode=590239,
+    text_x_uuencode=590240,
-    text_x_vcalendar=590240,
+    text_x_vcalendar=590241,
-    text_x_vcard=590241,
+    text_x_vcard=590242,
-    text_xml=590242 | 0x01000000,
+    text_xml=590243 | 0x01000000,
-    video_MP2T=393635,
+    video_MP2T=393636,
-    video_animaflex=393636,
+    video_animaflex=393637,
-    video_avi=393637,
+    video_avi=393638,
-    video_avs_video=393638,
+    video_avs_video=393639,
-    video_mp4=393639,
+    video_mp4=393640,
-    video_mpeg=393640,
+    video_mpeg=393641,
-    video_quicktime=393641,
+    video_quicktime=393642,
-    video_vdo=393642,
+    video_vdo=393643,
-    video_vivo=393643,
+    video_vivo=393644,
-    video_vnd_rn_realvideo=393644,
+    video_vnd_rn_realvideo=393645,
-    video_vosaic=393645,
+    video_vosaic=393646,
-    video_webm=393646,
+    video_webm=393647,
-    video_x_amt_demorun=393647,
+    video_x_amt_demorun=393648,
-    video_x_amt_showrun=393648,
+    video_x_amt_showrun=393649,
-    video_x_atomic3d_feature=393649,
+    video_x_atomic3d_feature=393650,
-    video_x_dl=393650,
+    video_x_dl=393651,
-    video_x_dv=393651,
+    video_x_dv=393652,
-    video_x_fli=393652,
+    video_x_fli=393653,
-    video_x_flv=393653,
+    video_x_flv=393654,
-    video_x_isvideo=393654,
+    video_x_isvideo=393655,
-    video_x_jng=393655 | 0x80000000,
+    video_x_jng=393656 | 0x80000000,
-    video_x_m4v=393656,
+    video_x_m4v=393657,
-    video_x_matroska=393657,
+    video_x_matroska=393658,
-    video_x_mng=393658,
+    video_x_mng=393659,
-    video_x_motion_jpeg=393659,
+    video_x_motion_jpeg=393660,
-    video_x_ms_asf=393660,
+    video_x_ms_asf=393661,
-    video_x_msvideo=393661,
+    video_x_msvideo=393662,
-    video_x_qtc=393662,
+    video_x_qtc=393663,
-    video_x_sgi_movie=393663,
+    video_x_sgi_movie=393664,
-    x_epoc_x_sisx_app=721344,
+    x_epoc_x_sisx_app=721345,
 };
 char *mime_get_mime_text(unsigned int mime_id) {switch (mime_id) {
 case application_arj: return "application/arj";
@@ -804,6 +805,7 @@ case text_mcf: return "text/mcf";
 case text_pascal: return "text/pascal";
 case text_PGP: return "text/PGP";
 case text_plain: return "text/plain";
+case application_vnd_coffeescript: return "application/vnd.coffeescript";
 case text_richtext: return "text/richtext";
 case text_rtf: return "text/rtf";
 case text_scriplet: return "text/scriplet";
@@ -1315,6 +1317,11 @@ g_hash_table_insert(ext_table, "sfv", (gpointer)text_plain);
 g_hash_table_insert(ext_table, "m3u", (gpointer)text_plain);
 g_hash_table_insert(ext_table, "csv", (gpointer)text_plain);
 g_hash_table_insert(ext_table, "eml", (gpointer)text_plain);
+g_hash_table_insert(ext_table, "make", (gpointer)text_plain);
+g_hash_table_insert(ext_table, "log", (gpointer)text_plain);
+g_hash_table_insert(ext_table, "markdown", (gpointer)text_plain);
+g_hash_table_insert(ext_table, "yaml", (gpointer)text_plain);
+g_hash_table_insert(ext_table, "coffee", (gpointer)application_vnd_coffeescript);
 g_hash_table_insert(ext_table, "rt", (gpointer)text_richtext);
 g_hash_table_insert(ext_table, "rtf", (gpointer)text_richtext);
 g_hash_table_insert(ext_table, "rtx", (gpointer)text_richtext);
@@ -1792,6 +1799,7 @@ g_hash_table_insert(mime_table, "text/mcf", (gpointer)text_mcf);
 g_hash_table_insert(mime_table, "text/pascal", (gpointer)text_pascal);
 g_hash_table_insert(mime_table, "text/PGP", (gpointer)text_PGP);
 g_hash_table_insert(mime_table, "text/plain", (gpointer)text_plain);
+g_hash_table_insert(mime_table, "application/vnd.coffeescript", (gpointer)application_vnd_coffeescript);
 g_hash_table_insert(mime_table, "text/richtext", (gpointer)text_richtext);
 g_hash_table_insert(mime_table, "text/rtf", (gpointer)text_rtf);
 g_hash_table_insert(mime_table, "text/scriplet", (gpointer)text_scriplet);

src/parsing/parse.c

@@ -38,6 +38,8 @@ void fs_reset(struct vfile *f) {
     }
 }
 
+#define IS_GIT_OBJ (strlen(doc.filepath + doc.base) == 38 && (strstr(doc.filepath, "objects") != NULL))
+
 void parse(void *arg) {
 
     parse_job_t *job = arg;
@@ -77,6 +79,10 @@ void parse(void *arg) {
     int bytes_read = 0;
 
     if (doc.mime == 0 && !ScanCtx.fast) {
+        if (IS_GIT_OBJ) {
+            goto abort;
+        }
+
         // Get mime type with libmagic
         if (!job->vfile.is_fs_file) {
             LOG_WARNING(job->filepath, "Guessing mime type with libmagic inside archive files is not currently supported");

src/static/css/dark.css

@@ -218,17 +218,25 @@ a:hover, .btn:hover {
     background-color: #e0e0e0;
 }
 
+.card-img-top {
+    border-top-left-radius: 0;
+    border-top-right-radius: 0;
+}
+
 .fit {
     display: block;
     min-width: 64px;
     max-width: 100%;
-    max-height: 240px;
+    max-height: 400px;
     margin: 0 auto 0;
-    padding: 3px 3px 0;
     width: auto;
     height: auto;
 }
 
+.img-padding {
+    padding: 4px 4px 0 4px;
+}
+
 .fit-sm {
     display: block;
     max-width: 64px;
@@ -530,3 +538,7 @@ svg {
 .stat > .card-body {
     padding: 0.7em 1.25em;
 }
+
+#modal-body > .img-wrapper {
+    margin-bottom: 1em;
+}

src/static/css/light.css

@@ -157,18 +157,25 @@ body {
     overflow: hidden;
 }
 
+.card-img-top {
+    border-top-left-radius: 0;
+    border-top-right-radius: 0;
+}
 
 .fit {
     display: block;
     min-width: 64px;
     max-width: 100%;
-    max-height: 240px;
+    max-height: 400px;
     margin: 0 auto 0;
-    padding: 3px 3px 0 3px;
     width: auto;
     height: auto;
 }
 
+.img-padding {
+    padding: 4px 4px 0 4px;
+}
+
 .fit-sm {
     display: block;
     max-width: 64px;
@@ -409,4 +416,8 @@ mark {
 
 .stat > .card-body {
     padding: 0.7em 1.25em;
+}
+
+#modal-body > .img-wrapper {
+    margin-bottom: 1em;
 }

src/static/js/dom.js

@@ -75,6 +75,7 @@ function shouldPlayVideo(hit) {
         hit["_source"]["extension"] !== "mkv" &&
         hit["_source"]["extension"] !== "avi" &&
         videoc !== "hevc" &&
+        videoc !== "mpeg1video" &&
         videoc !== "mpeg2video" &&
         videoc !== "wmv3";
 }
@@ -194,10 +195,21 @@ function makeUserTag(tag, hit) {
 function infoButtonCb(hit) {
     return () => {
         getDocumentInfo(hit["_id"]).then(doc => {
+            $("#modal-body").empty()
+
             $("#modal-title").text(doc["name"] + ext(hit));
 
+            if (doc["mime"]) {
+                const mimeCategory = doc["mime"].split("/")[0];
+                const imgWrapper = document.createElement("div");
+                imgWrapper.setAttribute("style", "position: relative");
+                imgWrapper.setAttribute("class", "img-wrapper");
+                makeThumbnail(mimeCategory, hit, imgWrapper, false);
+                $("#modal-body").append(imgWrapper);
+            }
+
             const tbody = $("<tbody>");
-            $("#modal-body").empty()
+            $("#modal-body")
                 .append($("<table class='table table-sm'>")
                     .append($("<thead>")
                         .append($("<tr>")
@@ -208,10 +220,19 @@ function infoButtonCb(hit) {
                     .append(tbody)
                 );
 
+            tbody.append($("<tr>")
+                .append($("<td>").text("index"))
+                .append($("<td>").text(`[${indexMap[doc["index"]]}]`))
+            ).append($("<tr>")
+                .append($("<td>").text("mtime"))
+                .append($("<td>")
+                    .text(new Date(doc["mtime"] * 1000).toISOString().split(".")[0].replace("T", " "))
+                    .attr("title", doc["mtime"]))
+            );
             const displayFields = new Set([
-                "mime", "size", "mtime", "path", "title", "width", "height", "duration", "audioc", "videoc",
+                "mime", "size", "path", "title", "width", "height", "duration", "audioc", "videoc",
                 "bitrate", "artist", "album", "album_artist", "genre", "title", "font_name", "tag", "author",
-                "modified_by"
+                "modified_by", "pages"
             ]);
             Object.keys(doc)
                 .filter(key => key.startsWith("_keyword.") || key.startsWith("_text.") || displayFields.has(key) || key.startsWith("exif_"))
@@ -392,7 +413,11 @@ function makeThumbnail(mimeCategory, hit, imgWrapper, small) {
     if (small) {
         thumbnail.setAttribute("class", "fit-sm");
     } else {
-        thumbnail.setAttribute("class", "card-img-top fit");
+        if (hit["_source"].hasOwnProperty("parent")) {
+            thumbnail.setAttribute("class", "card-img-top fit img-padding");
+        } else {
+            thumbnail.setAttribute("class", "card-img-top fit");
+        }
     }
     thumbnail.setAttribute("src", `t/${hit["_source"]["index"]}/${hit["_id"]}`);
 

src/static/search.html

@@ -12,7 +12,7 @@
 
 <nav class="navbar navbar-expand-lg">
     <a class="navbar-brand" href="/">sist2</a>
-    <span class="badge badge-pill version">2.7.4</span>
+    <span class="badge badge-pill version">2.8.4</span>
     <span class="tagline">Lightning-fast file system indexer and search tool </span>
     <a class="btn ml-auto" href="/stats">Stats</a>
     <button class="btn" type="button" data-toggle="modal" data-target="#settings" onclick="loadSettings()">Settings
@@ -227,7 +227,6 @@
                             <option value="7">7</option>
                             <option value="8">8</option>
                             <option value="9">9</option>
-                            <option value="13">13</option>
                         </select>
                     </div>
 

src/static/stats.html

@@ -10,7 +10,7 @@
 
 <nav class="navbar navbar-expand-lg">
     <a class="navbar-brand" href="/">sist2</a>
-    <span class="badge badge-pill version">2.7.4</span>
+    <span class="badge badge-pill version">2.8.4</span>
     <span class="tagline">Lightning-fast file system indexer and search tool </span>
     <a style="margin-left: auto" class="btn" href="/">Back</a>
     <button class="btn" type="button" data-toggle="modal" data-target="#settings"

src/web/serve.c

@@ -10,8 +10,6 @@
 
 #include <mongoose.h>
 
-#define CHUNK_SIZE 1024 * 1024 * 10
-
 
 static int has_prefix(const struct mg_str *str, const struct mg_str *prefix) {
     return str->len > prefix->len && memcmp(str->p, prefix->p, prefix->len) == 0;
@@ -237,13 +235,12 @@ void search(struct mg_connection *nc, struct http_message *hm) {
     *(body + hm->body.len) = '\0';
 
     char url[4096];
-    snprintf(url, 4096, "%s/sist2/_search", WebCtx.es_url);
+    snprintf(url, 4096, "%s/%s/_search", WebCtx.es_url, WebCtx.es_index);
 
     nc->user_data = web_post_async(url, body);
-    free(body);
 }
 
-int serve_file_from_url(cJSON *json, index_t *idx, struct mg_connection *nc) {
+void serve_file_from_url(cJSON *json, index_t *idx, struct mg_connection *nc) {
 
     const char *path = cJSON_GetObjectItem(json, "path")->valuestring;
     const char *name = cJSON_GetObjectItem(json, "name")->valuestring;
@@ -538,8 +535,8 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
             }
         }
 
-        char buf[8192];
+        char *buf = malloc(sizeof(char) * 8192);
-        snprintf(buf, sizeof(buf),
+        snprintf(buf, 8192,
                  "{"
                  "    \"script\" : {"
                  "        \"source\": \"if (ctx._source.tag.contains(params.tag)) { ctx._source.tag.remove(ctx._source.tag.indexOf(params.tag)) }\","
@@ -552,14 +549,14 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
         );
 
         char url[4096];
-        snprintf(url, sizeof(url), "%s/sist2/_update/%s", WebCtx.es_url, arg_req->doc_id);
+        snprintf(url, sizeof(url), "%s/%s/_update/%s", WebCtx.es_url, WebCtx.es_index, arg_req->doc_id);
         nc->user_data = web_post_async(url, buf);
 
     } else {
         cJSON_AddItemToArray(arr, cJSON_CreateString(arg_req->name));
 
-        char buf[8192];
+        char *buf = malloc(sizeof(char) * 8192);
-        snprintf(buf, sizeof(buf),
+        snprintf(buf, 8192,
                  "{"
                  "    \"script\" : {"
                  "        \"source\": \"if(ctx._source.tag == null) {ctx._source.tag = new ArrayList()} ctx._source.tag.add(params.tag)\","
@@ -572,7 +569,7 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
         );
 
         char url[4096];
-        snprintf(url, sizeof(url), "%s/sist2/_update/%s", WebCtx.es_url, arg_req->doc_id);
+        snprintf(url, sizeof(url), "%s/%s/_update/%s", WebCtx.es_url, WebCtx.es_index, arg_req->doc_id);
         nc->user_data = web_post_async(url, buf);
     }
 
@@ -668,11 +665,11 @@ static void ev_router(struct mg_connection *nc, int ev, void *p) {
         if (nc->user_data != NULL) {
             //Waiting for ES reply
             subreq_ctx_t *ctx = (subreq_ctx_t *) nc->user_data;
-            mg_mgr_poll(&ctx->mgr, 0);
+            web_post_async_poll(ctx);
 
-            if (ctx->ev_data.done == TRUE) {
+            if (ctx->done == TRUE) {
 
-                response_t *r = ctx->ev_data.resp;
+                response_t *r = ctx->response;
 
                 if (r->status_code == 200) {
                     send_response_line(nc, 200, r->size, "Content-Type: application/json");
@@ -695,6 +692,8 @@ static void ev_router(struct mg_connection *nc, int ev, void *p) {
                 }
 
                 free_response(r);
+                free(ctx->data);
+                free(ctx);
                 nc->flags |= MG_F_SEND_AND_CLOSE;
                 nc->user_data = NULL;
             }