Unicodes Gone Wild
Talkat AppSec California 2015 · on Tuesday, 2015-01-27· at 13:45· Annenberg Community Beach House: Track 1, 415 Pacific Coast Hwy·
summary
Explains concrete Unicode failures and measurements: notes 86.4% of web pages use UTF-8, demonstrates differences between UTF-8/16/32 and how Python/Ruby internal widths change string lengths, uses Charles proxy to show servers returning 500/503 when fed long combining sequences, builds a 256‑combining‑character filename that makes Google Drive crash for ~30 seconds and refuse deletion until renamed, and reproduces a Spotify username normalization lets reset hijack scenario that actually succeeds.
AI-written summary · about
chapters (9)
- 0:00Introduction to Unicode
- 3:06History and standard evolution
- 7:22Technical structure
- 11:21Encoding formats
- 16:49Complex character features
- 24:38Operational challenges
- 29:08Compiler and language behavior
- 38:54Security implications
- 50:48Best practices and conclusion
