| Summary: | pdf file crash | ||
|---|---|---|---|
| Product: | MuPDF | Reporter: | sylvain.bertrand |
| Component: | mupdf | Assignee: | Tor Andersson <tor.andersson> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | ||
| Priority: | P4 | ||
| Version: | unspecified | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| Customer: | Word Size: | --- | |
| Attachments: | guilty file | ||
|
Description
sylvain.bertrand
2010-09-01 10:06:37 UTC
Created attachment 6684 [details]
guilty file
The file has an inline run-length encoded image, with garbage on the end: a '>' character. The '>' character is used as the end-of-data marker for ASCIIHex and ASCII85 decoding filters, but not for RunLength -- perhaps the generator is confused about the spec? I have added a workaround that ignores garbage at the end of inline images and scans for the EI keyword without using the lexer. damn... the pdf format has too many damaged producers, need an official validation tool. thx for the help. |