Hey so I just realized that if you try to have several subsequent calls to the http derivation, and because you don’t know the hash, you naively try to put the same each time, it will fail in a surprising way:
the first http will download and store at the right place (provided it’s the right hash)
the second will look for a file with the same name and find it (it’s the first)
elpe will check for the hash and say it’s correct
Everything will be OK, except it isn’t.
I don’t see any good workaround for this, as encoding the url name in the file name would probably not be desirable (it’s OK if the url changes but not the content, for example to deal with tokens in urls).
I haven’t tried that with nix, I don’t know how it’s handled there.
Hey so I just realized that if you try to have several subsequent calls to the http derivation, and because you don’t know the hash, you naively try to put the same each time, it will fail in a surprising way:
I don’t see any good workaround for this, as encoding the url name in the file name would probably not be desirable (it’s OK if the url changes but not the content, for example to deal with tokens in urls).
I haven’t tried that with nix, I don’t know how it’s handled there.
Madx