Drop Adobe Illustrator (.ai) detection support#743
Conversation
|
Agreed, but this is a breaking change, and we just did a major version, so I think this should wait a bit. |
|
You could maybe consider AI detection for your XML detection library as I think I remember that you can detect Illustrator files by looking at its XMP metadata, which is XML. |
No problem, I am not in a rush.
Something like that. Starting with basic PDF decoding which I could maybe utilize read-next-line to iterate over the lines. I am not very familiar with the PDF file format though. |
|
Adobe Illustrator ( |
Drop Adobe Illustrator (
.ai) file detection support as the detection mechanism is very poor quality (as raised in #582).Skipping 1350 bytes, followed by a search in the next 10kB for a keywords is not a reliable mechanism.
There is no fixed offset at 1350 bytes, as well the 10kB is not based on any meaningful foundation.
And neither the keywords necessary from from a context which reliably indicates this is an Adobe Illustrator file.
Second objection I have, is that this is based on (poor) text file parsing, as PDF is a text based format.
Text based formats are not in the scope of file-type.
Reverses: #323
Resolves: #582
Related: