-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Jochen Roderburg wrote:
(abbreviated:) > 51% [=============================> ] 157,962 > 5.37K/s in 3m 45s > Further examination of such cases showed that a Byte-Range transfer was > requested by wget, but was not done by the server. On most normal cases this > cannot happen, because wget first examines the headers and does these requests > only when they are advertized by the server. The situation, where this does > happen nevertheless, is very complicated and rare and envolves local and > remote > proxies again. Thanks for tracking this down, Jochen. Probably not something to worry too much about for 1.11, then; I've created the bug report and targeted it for 1.12. I'm still waiting to have a chance to have a good look at the other stuff you brought up (related to HEAD, timestamping, etc). I've decided that it's probably worth a potential delay (assuming that the legal stuff gets resolved soon) to get some solid, machine-interpretable automated tests on this behavior, so we have a clear understanding of what Wget is doing in that area. AIUI, the tests Mauro wrote for that code are intended for human consumption (that is, look over the output and verify that it's doing the right thing), which isn't practical enough for regular regression testing, because our current test framework doesn't support protocol-level testing very well (though, with a bit of tweaking, I believe it can be brought to an adequate level for that). To all: see http://wget.addictivecode.org/FeatureSpecifications/Testing for thoughts on where testing should maybe go in the future. - -- Micah J. Cowan Programmer, musician, typesetting enthusiast, gamer... http://micah.cowan.name/ -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFG+EH47M8hyUobTrERCA0OAJ4i8jQYPmmgqEgH8JfOgdwCPK0xawCfYwHM X+xOZ3eJ8cd29Jr2piJo0HU= =+LGc -----END PGP SIGNATURE-----
