-- Mat Garrett [EMAIL PROTECTED]
--- Begin Message --- I have some suggestions for small modifications for enhancing future versions of wget.
I understand the need to first retrieve a .listing when using --timestamping with an ftp:// URL, but shouldn't that step be skipped when downloading the file for the first time? I.e., couldn't you do a check for the existence of the downloaded file, and if it doesn't exist, ignore the --timestamping argument and go straight to retrieving the file itself.
When I use "wget --timestamping --output-document foo url://site.com/bar", shouldn't wget check the size and mtime of `bar' against `foo'? I seem to be getting aberant behaviour in that even if `foo' is a properly up-to-date copy of `bar', it still gets retrieved again.
I'm currently in the middle of a project to reinvent the wheel, where wheel equals portage/LSM/SlackBuild/RPM .spec files/et al. One aspect of my package management system is the automated retrieval of all required package files. Every part of my PM system is handled by an agent abstraction. So far, the only download-agent I've felt the need to write is the wrapper around wget. It just kicks that much butt. I'm not looking forward to writing the various and sundry revision control system download-agents, though.
Because maximum flexibility is desired, I'm not pooling together all retrievals that come out of the same repository, each retrieval is treated in isolation. For the most part, this is not a problem, but for things like EMACS, where there are a number of files, all in the same repository, it'd be convenient to be able to cache the .listing file on the first retrieval, and then have wget reuse the same .listing for each subsequent retrieval from the corresponding repository. For your part, all it should take would be adding something akin to the cookies args for listings: --save-listing <listing file>, --use-listing <listing file>. All this is intended to do is prevent a small handful of redundent ftp listing retrievals, so it has a very low priority for me.
Thank you for producing such a great utility!
I've used --random-wait, --referer=, and --user-agent="Internet Explorer" many times to get around idiotic http server restrictions, but I've never found a single resource I could get to through a regular browser that wget couldn't get at as well, even when using forms and cookies.
-- Mat Garrett [EMAIL PROTECTED]
--- End Message ---
