wallabag icon indicating copy to clipboard operation
wallabag copied to clipboard

wallabag is not displaying base64 encoded images

Open HolgerAusB opened this issue 1 year ago • 2 comments

Environment

  • Version: 2.6.10
  • Installation: make install/update
  • PHP version: 8.2.26
  • OS: Debian Trixie=testing
  • Database: mysql
  • Parameters:
My app/config/parameters.yml is:
parameters:
  database_driver: pdo_mysql
  database_host: 127.0.0.1
  database_port: 3306
  database_name: xxx
  database_user: xxx
  database_password: xxxx
  database_path: null
  database_table_prefix: xxxx_
  database_socket: null
  database_charset: utf8mb4
  domain_name: 'https://xxx.example.com'
  locale: de
  secret: xxx
  twofactor_sender: xxx@xxx
  fosuser_registration: false
  fosuser_confirmation: true
  from_email: xxx@xxx
  rss_limit: 50
  rabbitmq_host: localhost
  rabbitmq_port: 5672
  rabbitmq_user: xxx
  rabbitmq_password: xxx
  rabbitmq_prefetch_count: 10
  redis_scheme: tcp
  redis_host: localhost
  redis_port: 6379
  redis_path: null
  redis_password: null
  server_name: xxx.example.com
  fos_oauth_server_access_token_lifetime: 3600
  fos_oauth_server_refresh_token_lifetime: 1209600
  sentry_dsn: null
  mailer_dsn: 'smtp://127.0.0.1'

What steps will reproduce the bug?

Trying to catch

  • https://papertohtml.org/paper?id=d4685c38dbf463a0d6398eafb9944a056af3b0a8
  • https://immub.org/noticias/tambor-de-crioula-do-piaui-a-africa-brasileira while the second one is a JavaScript re-direct to
  • https://pub-24ea3407a0f24c8394eadaa45a05aa0c.r2.dev/texto/3984_pt.html?update=ILT

All images are NOT linked with <img src="http... but directly encoded as in-html base64 images. While FTR could display the images with current site-configs, in wallabag I only see a verrrrrry long literal <img src="data:image/png;base64,iVBO...

Screenshot of wallabag UI after the fetch: grafik

HolgerAusB avatar Dec 26 '24 12:12 HolgerAusB

just tested some string_replace on this new URL from another website, none of the following worked at f43.me, which is weird:

replace_string(<img): <foobar
replace_string(src): foobar
#first words from the real text:
replace_string(Special thanks to): foobar

I always get the same result, there is no 'foobar' in the in the excerpt, neither in html node/attribute nor in literal text.

@j0k3r please have a look on this, or should I open an issue in Graby repo?

HolgerAusB avatar Jan 26 '25 22:01 HolgerAusB

+1 I'm having the same issue - used single file to workaround download the full page of a site that has 2FA. The downloaded page have all image converted to base64 & when I send it over the wallabagger extension, the article did not render images but a long base64 string.

luutuankiet avatar Mar 26 '25 13:03 luutuankiet