Re: [Bug-wget] WARC, new version

[Giuseppe Scrivano]

Gijs van Tulder <address@hidden> writes:

> Hi all,
>
> Based on the comments by Giuseppe and Ángel I've revised the
> implementation of the wget WARC extenstion. I've attached a patch.
>
> 1. It's no longer based on the warctools library. Instead, I've
> written a couple of new WARC-writing functions, using zlib for the
> gzip compression. The new implementation is much smaller.
>
> 2. I extracted a small part of the gethttp method in http.c and moved
> it to a new function, read_response_body, which is responsible for
> downloading the response body and writing it to a file.
>
> The WARC extension needs to save the response in multiple cases: when
> the response is successful, but also when the response is a redirect,
> 401 unauthorized or an error. Moving the response-saving to a separate
> method makes it possible to reuse this part for all four situations.
>
> Any thoughts?

WOW great work!  It is much better now.

I wonder if it is possible to remove the dependency from libuuid, maybe
provide replacement for uuid_generate and uuid_unparse when libuuid is
not found?  Even a simple implementation based on rand?

Beside it, there are only very small adjustments which need to be done
to the code in order to include it into wget, like lines not longer than
80 characters or using "foo *bar" instead of "foo * bar"; in any case
these are not important and I can go trough them before commit your
changes.

Thanks,
Giuseppe