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

Re: Using w3m-download



I've posted to my github fork a working version of a function
`w3m-download-using-wget' which works real neatly, and has an added
bonus of optionally tagging png and jpg files with their captions
(requires other external programs). The reason I haven't yet issued a
pull request for it is that it doesn't yet ever use the emacs-w3m cache
for files that might already be locally stored. If the project finds the
work desirable, the next question would be how to integrate it into the
code-base. As a separate command? As a replacement? In combination with
the existing command?

The work is available as a stand-alone download at:

https://github.com/Boruch-Baum/emacs-w3m/raw/bb_download/w3m-download.el

So, anyone on the list can easily try it out without doing a git clone
etc.

Just a few notes:

1] Because it's using wget, downloads are resumable! If a download is
   aborted for any reason, just run the command again.

2] C-u allows you to rename the output file.

3] Each download has its own unique progress buffer *<download[n]>*. One
   way to abort a download is to kill its progress buffer. Another  way
   is to use `pgrep -a` and `kill` on the command line, but emacs now
   has its own internal version of `top`, so you can run `M-x
   list-processes', move point to a process that you want to kill, and
   press `d'.

All feedback welcome.


On 2019-03-07 21:23, Boruch Baum wrote:
> Instead of using wget, I decided to try emacs-w3m for downloading
> several large (~500Mb) files, and after crashing my emacs server, have
> a few comments:
>
> 1) function w3m-download uses function w3m-retrieve, which says it will
>    put the data into the current buffer. Is that ever desirable for a
>    download? I performed an lgrep to see how the project uses
>    w3m-download and even for small uses, it doesn't ever seem necessary,
>    optimal, or even desirable.
>
> 2) There is no progress indicator of any kind.
>
> 3) I thought maybe I could find the partial downloads in some temporary
>    files somewhere, but didn't see any, not in ~/.w3m or ~/.emacs.d.
>
> 4) There's no way to resume an aborted download.
>
> Am I correct about all this? Is the project interested in these
> features? If these features aren't available with the w3m back-end, is
> there an objection to using something else, such as wget or aria2?
>
> I've coded something similar for another project, so I can share the
> generic code.
>

-- 
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0