Brad Wardman wrote:
Micah,
I forgot to add that I am using the -r (recursive flag)
wget version 1.10.2 (Red Hat modified)
CentOS 5.3 - Linux 2.6.18 kernel
Ah, okay. I had a hunch that was the case.
Unfortunately, Red Hat's (CentOS's) wget 1.10.2 diverged quite a bit
from ours, and slurped in quite a few unstable features from 1.11 while
it was still in development/unstable. It probably included
Content-Disposition support, but from a time when it was on by default,
and had (more) directory-placement bugs.
My suggestion would be to grab an updated copy of wget (RPMs of 1.11.4
or 1.12 should be fine: I don't believe they have been nearly as
customized as 1.10.2 was), or else to build 1.10.2 from the official GNU
sources.
As a cheap workaround, --no-content-disposition or "-e
content_disposition=off" may disable content-disposition support in
wget, thus avoiding the directory-placement bug. However, your .pdf
files will show up as "index.php?blahblahblah" files, and won't have the
.pdf extensions. Likewise, if you grab a 1.11.4 or 1.12 (regardless of
whether they're Red Hat's or GNU's), you'd need to _add_
--content-disposition to get that renaming action (note however that
--content-disposition can sometimes result in extra round-trips to the
server to get information about URLs, which is why it's not enabled by
default. But you were hitting those inefficiencies already with your Red
Hat-modified wget 1.10.2, so if it wasn't bothering you before, it won't
bother you now).