Hi,
On Tue, Feb 7, 2017 at 4:56 PM, Adrien Destugues <pulkomandy@xxxxxxxxxxxxx>
wrote:
7 février 2017 16:45 "Dario Casalinuovo" <b.vitruvio@xxxxxxxxx> a écrit:
* Binary and textual files are distinguished from start being one or theother category (this has
nothing to do with the filetype section though)
Do you handle various text file encodings? Because not everything text
looks like ASCII (UTF16, EBCDIC, ...).
* Binary files are recognized by the magic number or defaults to genericbinary type otherwise.
Our MIME rules can scan up to the first 512 bytes of the file, and even
that is sometimes barely enough to detect the format. Our sniffing rules
can do things a little more flexible than just recognizing a fixed magic
number.
* Textual files are recognized using prebuilt (made by human or machinegenerated) rules or
defaults to plain text otherwise.LGPL), so I'm wondering, is
I have the freedom to release it as an open source library (possibly
there any interest having something like that integrated in Haiku?and Haiku.
As long as I'm in an early stage I can make things works on both linux
Why not?
I don't know from your short description how far you plan to go. If it
does as good as our existing MIME sniffer or better, and it is as fast, we
would definitely replace the MIME sniffer with something more powerful. The
transition would be easier if there is some tool to convert our existing
MIME rules to the new format, however. Or maybe we can make something where
the MIME sniffing can be made by several add-ons, and the existing MIME
sniffer and your new tool can collaborate?