Add limit to HTTP service's page size

Fixes #50
browsh-org · Jun 23, 2018 · 9ce338f · 9ce338f
1 parent 61626b5
commit 9ce338f
Show file tree

Hide file tree

Showing 2 changed files with 20 additions and 2 deletions.
diff --git a/webext/src/dom/dimensions.js b/webext/src/dom/dimensions.js
@@ -22,6 +22,11 @@ export default class extends utils.mixins(CommonMixin) {
     // sent to the TTY to be buffered to support faster scrolling.
     this._big_sub_frame_factor = 6;
 
+    // The max size in pixels for either the width or height to be for Browsh to parse in
+    // raw text mode.
+    // TODO: Use incremental parses to overcome this limit.
+    this._entire_dom_limit = 30000;
+
     this.dom = {};
     this.tty = {};
     this.frame = {
@@ -100,6 +105,14 @@ export default class extends utils.mixins(CommonMixin) {
       width: this.dom.width,
       height: this.dom.height,
     }
+    if (this.dom.sub.width > this._entire_dom_limit) {
+      this.dom.sub.width = this._entire_dom_limit;
+      this.is_page_truncated = true;
+    }
+    if (this.dom.sub.height > this._entire_dom_limit) {
+      this.dom.sub.height = this._entire_dom_limit;
+      this.is_page_truncated = true;
+    }
     this.frame.sub = {
       left: 0,
       top: 0,

diff --git a/webext/src/dom/serialise_mixin.js b/webext/src/dom/serialise_mixin.js
@@ -27,7 +27,8 @@ export default (MixinBase) => class extends MixinBase {
   }
 
   _serialiseRawText() {
-    let raw_text = "";
+    let info = '';
+    let raw_text = '';
     this._previous_cell_href = '';
     this._is_inside_anchor = false;
     const top = this.dimensions.frame.sub.top / 2;
@@ -41,7 +42,11 @@ export default (MixinBase) => class extends MixinBase {
       raw_text += "\n";
     }
     const head = `<html><title>${document.title}</title><body><pre>`
-    const foot = `</body></pre></html>`
+    info += "\n\n" + 'Built by <a href="https://www.brow.sh">Browsh</a>'
+    if (this.dimensions.is_page_truncated) {
+      info += '\nBrowsh parser: the page was too large, some text may have been truncated.';
+    }
+    const foot = `${info}</pre></body></html>`
     return head + raw_text + foot;
   }