chore(chrome): add to block list

spider-rs · Jan 15, 2025 · 9ad8e4d · 9ad8e4d
1 parent ff770ce
commit 9ad8e4d
Show file tree

Hide file tree

Showing 10 changed files with 26 additions and 13 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/spider/Cargo.toml b/spider/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider"
-version = "2.26.0"
+version = "2.26.1"
 authors = [
     "j-mendez <[email protected]>"
 ]

diff --git a/spider/src/page.rs b/spider/src/page.rs
@@ -489,7 +489,6 @@ fn get_error_status(
     get_error_status_base(should_retry, error_for_status).map(std::sync::Arc::new)
 }
 
-
 /// Instantiate a new page without scraping it and with the base URL parsed (used for testing purposes).
 #[cfg(not(feature = "decentralized"))]
 pub fn build_with_parse(url: &str, res: PageResponse) -> Page {

diff --git a/spider_chrome/Cargo.toml b/spider_chrome/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider_chrome"
-version = "2.26.0"
+version = "2.26.1"
 rust-version = "1.70"
 authors = [
     "j-mendez <[email protected]>"

diff --git a/spider_chrome/src/handler/blockers/scripts.rs b/spider_chrome/src/handler/blockers/scripts.rs
@@ -68,6 +68,16 @@ lazy_static::lazy_static! {
             "https://acdn.adnxs.com/ast/ast.js",
             "https://schibsted-cdn.relevant-digital.com/static/tags/",
             "https://bat.bing.net",
+            "https://tpc.googlesyndication.com/",
+            "https://cdn.petametrics.com/",
+            "https://cdn.doubleverify.com/",
+            "https://www.facebook.com/v17.0/plugins/like.php?",
+            "https://gum.criteo.com",
+            "https://js-sec.indexww.com",
+            "https://eus.rubiconproject.com/",
+            "https://eb2.3lift.com/",
+            "https://acdn.adnxs.com/",
+            "https://ssc-cms.33across.com/",
             "https://static.addtoany.com/menu/",
             "https://www.gstatic.com/cast/sdk/libs/sender/1.0/cast_framework.js",
             "https://www.gstatic.com/eureka/clank/131/cast_sender.js",
@@ -110,6 +120,7 @@ lazy_static::lazy_static! {
             "https://ads.pubmatic.com/AdServer/js/",
             "https://widgets.outbrain.com/nanoWidget/externals/obPixelFrame/obPixelFrame.js",
             "https://widgets.outbrain.com/external/externals/intentiq.js",
+            "https://cdn.fuseplatform.net/publift/tags/",
             "//d2wy8f7a9ursnm.cloudfront.net/v8/bugsnag.min.js",
             ".sharethis.com",
             ".newrelic.com",

diff --git a/spider_chrome/src/handler/blockers/xhr.rs b/spider_chrome/src/handler/blockers/xhr.rs
@@ -86,9 +86,12 @@ lazy_static::lazy_static! {
             "https://sdkconfig.pulse.",
             "https://bat.bing.net",
             "https://api.reviews.io/",
+            "https://thearenagroup.sp.spiny.ai/",
             "https://ads.rubiconproject.com/",
+            "https://check.analytics.rlcdn.com",
             "https://api.config-security.com/event",
             "https://conf.config-security.com/model",
+            "https://pagead2.googlesyndication.com/",
             "https://sumome.com/api/load/",
             "https://public-api.wordpress.com/geo/",
             "https://events.api.secureserver.net/",

diff --git a/spider_cli/Cargo.toml b/spider_cli/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider_cli"
-version = "2.26.0"
+version = "2.26.1"
 authors = [
     "j-mendez <[email protected]>"
 ]

diff --git a/spider_transformations/Cargo.toml b/spider_transformations/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider_transformations"
-version = "2.26.0"
+version = "2.26.1"
 authors = [
     "j-mendez <[email protected]>"
 ]

diff --git a/spider_utils/Cargo.toml b/spider_utils/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider_utils"
-version = "2.26.0"
+version = "2.26.1"
 authors = [
     "j-mendez <[email protected]>"
 ]

diff --git a/spider_worker/Cargo.toml b/spider_worker/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "spider_worker"
-version = "2.26.0"
+version = "2.26.1"
 authors = [
     "j-mendez <[email protected]>"
 ]