Remove null chars, i.e. \u0000, when getting all text (PR 16286 follow-up)

I was playing with the new "copy all text" feature, and stumbled upon one document where the copied text was truncated; see http://mirrors.ctan.org/info/lshort/english/lshort.pdf

The problem turns out to be that on [page 83](https://ftp.acc.umu.se/mirror/CTAN/info/lshort/english/lshort.pdf#page=83) the textLayer contains `\u0000` and apparently copying just stops when a null char is encountered.
To fix this we can simply use an existing helper function, and with this patch we're able to successfully copy all the text in that document.

web/pdf_viewer.js

@@ -47,6 +47,7 @@ import {
   MAX_SCALE,
   MIN_SCALE,
   PresentationModeState,
+  removeNullCharacters,
   RenderingStates,
   SCROLLBAR_PADDING,
   scrollIntoView,
@@ -671,7 +672,7 @@ class PDFViewer {
           buffer.push("\n");
         }
       }
-      texts.push(buffer.join(""));
+      texts.push(removeNullCharacters(buffer.join("")));
     }
 
     return texts.join("\n");