This issue can (currently) be reproduced with ruby-highline.
Although now available (only) in community, the PKGBUILD still resides on the aur server -- this tricks aurget into thinking it's an aur package and therefore fails trying to install it.
If you suspect you're bit by this issue, please note the following which identifies these packages:
$ aurget -Ssi <package> # will find nothing, but
$ aurget -Sp <package> # will still print a PKGBUILD
The reason for this issue is that aurget attempts to locate and source the PKGBUILD during dependency resolution. Branching off of a 404 at that point is the most efficient way to know if something is available in the aur or (potentially) available via pacman. Solving this issue would mean adding additional verification for one of those at that point (something which already must occur at a later point in the process anyway). This would mean a dip in efficiency just to catch outliers like these.
I will leave this issue open for others to find until a) I come up with an efficient and simple way to code around it, b) the aur is fixed such that these PKGBUILDs are no longer left or c) I move this note to the README and leave things as-is.
This issue can (currently) be reproduced with ruby-highline.
Although now available (only) in community, the PKGBUILD still resides on the aur server -- this tricks aurget into thinking it's an aur package and therefore fails trying to install it.
If you suspect you're bit by this issue, please note the following which identifies these packages:
The reason for this issue is that aurget attempts to locate and source the PKGBUILD during dependency resolution. Branching off of a 404 at that point is the most efficient way to know if something is available in the aur or (potentially) available via pacman. Solving this issue would mean adding additional verification for one of those at that point (something which already must occur at a later point in the process anyway). This would mean a dip in efficiency just to catch outliers like these.
I will leave this issue open for others to find until a) I come up with an efficient and simple way to code around it, b) the aur is fixed such that these PKGBUILDs are no longer left or c) I move this note to the README and leave things as-is.