Expander: use mime type/sniffer to detect archive type as fallback

[umccullough]

It seems that I have downloaded a .tgz file that has the wrong extension (it's actually a .tar). Someone told me this was possibly a Web+ bug, but that doesn't matter.

What I'd like is for Expander to not fail trying to extract this file, but to recognize it's really a .tar file instead. It seems the mime sniffer has figured this out, so it would be nice if that was the fallback when no files were located in the archive based on the extension.

The file in question is the tuxpaint archive downloaded from here using Web+:

​http://zeta-games.com/remository/start-download/13-educational/113-tuxpaint-0.9.21-haiku.html

Note, when I download it with Chrome on Windows, it does appear to be a proper .tgz, but with Web+ on Haiku, I have to rename it .tar (or eliminate the extension entirely) before Expander can unarchive it.

[phoudoin]

Regarding to any zip or gzip file downloaded via WebPositive, it will be corrupted because in fact the file is on-the-fly decompressed by the curl network backend Web+ currently use.

This behavior is governed by CURLOPT_HTTP_CONTENT_DECODING and CURLOPT_HTTP_TRANSFER_DECODING options, which currently are not set to FALSE.

[axeld]

Only the latter really sounds like the option to change; the HTTP content should be decoded, or maybe WebKit has support for this as well itself.