ssokolow: The new
--cover option breaks my convenience solution for shell aliasing because it seems to provide no way to set cover=true and then override that in the few shell aliases which are the excepton to the rule.
(Instead, I have to write a shell function which checks for existing covers, backs them up if present, runs lgogdownloader, deletes newly-downloaded covers, and then restores any backups it may have made.)
369e8bb Add --no-cover option back
ssokolow: Also, I've ignored options like
--repair until now but, with the Linux release of The Witcher 2 being a single tarball that's just under 20GiB in size (which will take almost 24 hours to download here), I'd really appreciate instructions on what set of options to pass to lgogdownloader to request "Download the requested things, creating new files and folders as necessary, but re-download as little data as possible when encounting incomplete or corrupted files and do it in a way which is compatible with Ctrl+C or a power failure interrupting LGOGDownloader's previous attempt.")
...in fact, to be honest, having that as the default behaviour is what any employably competent UI/UX designer would insist on.
I know it's not
--repair --download because I seem to remember that being over-eager to re-download entire files and I know it's not
--repair because that gives a 404 error when asked to download something new.
Unfortunately there isn't a good way to do this other than running the downloader twice. First with --download to resume everything and then a second time with --repair to check and fix any possible corruptions.
This is a feature that I've been meaning to add to the dowloader for a while now but it's pretty painful to add because of the current structure of the downloader. This is one of the reasons I would like to restructure/rewrite the code.
I should have restructured the code a long time ago when I started adding more features and noticed that the current structure is bad. There's lots of duplicated code everywhere and some parts are really confusing due to multiple checks to accommodate all the new features.
However with Galaxy coming sometime soon (maybe?) and not knowing what happens to the old API, I don't really want to spend time on a rewrite just to realize that all the work was for nothing if/when the API gets discontinued.