[Date Prev][Date Next][Thread Prev][][Date Index][Thread Index]

Re: [Q] does w3m sometimes download garbage?



>> On Fri, 15 Oct 2004 10:34:42 +0900
>> yamaoka@jpl.org (Katsumi Yamaoka) said as follows:

>w3m-download specifies the second argument `no-decode' as t to the
>w3m-retrieve function so that it should not decode the contents.
>However, the contents of that url has been encoded with `deflate'.

>Does anyone know what is the right solution?

Short version of my answer:
  I think that Katsumi's proposal is the right solution.  `no-decode'
  argument specifies that `w3m-retrieve' should not decode contents
  even if they decoded with gzip, bzip2, or deflate.

Long version of my answer:
  There are three layers to encode contents in HTTP, as far as I know.
  The first layer is a filter to convert contents to fit transfer
  layer constraints including base64 encoder, and is specified by
  `Content-Transfer-Encoding' header.  The second layer is a archiver
  including gzip, bzip2 and deflate, and is specified by
  `Content-Encoding' header.  The last layer is an encoding of
  characters, and is specified by `charset' attribute in
  `Content-Type' header.

  `no-decode' argument of `w3m-retrieve' is corresponding to the
  second layer, and it specifies whether `w3m-retrieve' should expand
  compressed contents or not.

  If you want to display contents including other characters than
  ASCII, you will have to call `w3m-decode-buffer' in addition to
  calling `w3m-retrieve' without `no-decode' argument.

-- 
TSUCHIYA Masatoshi