Colon in the package version string denotes epoch, and is not present in the actual .deb filename.
For example, just on a normal system:
$ apt-get download bsdutils
Get:1 Downloading bsdutils 1:2.20.1-5.1ubuntu8 [40.4 kB]
Fetched 40.4 kB in 1s (32.8 kB/s)
$ ls
bsdutils_2.20.1-5.1ubuntu8_amd64.deb
But this does suggest that something in the installer relies on exact string matching, somewhere...
A simple test between GNU and busybox, does show differences in the output:
$ wget http://127.0.0.1/:
--2013-07-09 10:45:24-- http://127.0.0.1/:
Connecting to 127.0.0.1:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2013-07-09 10:45:24 ERROR 404: Not Found.
$ busybox wget http://127.0.0.1/:
Connecting to 127.0.0.1 (127.0.0.1:80)
wget: server returned error: HTTP/1.1 404 Not Found
Now looking into debian-installer-utils, reveals README.wget404 which explains in detail that fetching urls, depends on text matching of wget's output.
Have you considered adding a new fetch-url-methods/https, which uses wget.gnu (if available) with correct wget.gnu 404 parsing?
Or one can modify existing fetch-url-methods/http to accommodate for either busybox or GNU.
Colon in the package version string denotes epoch, and is not present in the actual .deb filename. 2.20.1- 5.1ubuntu8_ amd64.deb
For example, just on a normal system:
$ apt-get download bsdutils
Get:1 Downloading bsdutils 1:2.20.1-5.1ubuntu8 [40.4 kB]
Fetched 40.4 kB in 1s (32.8 kB/s)
$ ls
bsdutils_
But this does suggest that something in the installer relies on exact string matching, somewhere...
A simple test between GNU and busybox, does show differences in the output: 127.0.0. 1/: 127.0.0. 1/:
$ wget http://
--2013-07-09 10:45:24-- http://
Connecting to 127.0.0.1:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2013-07-09 10:45:24 ERROR 404: Not Found.
$ busybox wget http:// 127.0.0. 1/:
Connecting to 127.0.0.1 (127.0.0.1:80)
wget: server returned error: HTTP/1.1 404 Not Found
Now looking into debian- installer- utils, reveals README.wget404 which explains in detail that fetching urls, depends on text matching of wget's output.
Have you considered adding a new fetch-url- methods/ https, which uses wget.gnu (if available) with correct wget.gnu 404 parsing? methods/ http to accommodate for either busybox or GNU.
Or one can modify existing fetch-url-