The main reason I wrote the patch - I cannot build gnutls@3.8.3 from source code. This command ``` guix build gnutls@3.8.3 --no-substitutes ``` returns me this ``` expected hash: 0ghpyhhfa3nsraph6dws50jb3dc8g2cfl7dizdnyrm179fawakzp actual hash: 0kayjxy3rr2y08jjimz5f0dx92pq3xjiaj2pdnsn15h1rp2k21pa hash mismatch for store item '/gnu/store/9zh4mcmwp2afyl4ig56943w64czdz2kp-gnutls-3.8.3.tar.xz' ``` Guix I'm using ``` guix: repository URL: https://git.savannah.gnu.org/git/guix.git branch: master commit: 4009d1de954d694cb11af391d4113d29c5c1379d ``` Am I getting it right - you are suggesting I replace "mirror://gnupg/gnutls/v" with "https://gnutls.org/" or "ftp://gnutls.org" ? I'm asking because I'm not the original author of this package, so it's not obvious to me, what will be the right decision here? On Tue, Oct 29, 2024 at 8:59 PM Tomas Volf <~@wolfsden.cz> wrote: > > Adam Quandour writes: > > > Actual hash of this package is not the same as the specified one. > > --- > > gnu/packages/tls.scm | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/gnu/packages/tls.scm b/gnu/packages/tls.scm > > index 1a1ce0d..4f862ff 100644 > > --- a/gnu/packages/tls.scm > > +++ b/gnu/packages/tls.scm > > @@ -212,7 +212,7 @@ (define-public gnutls > > (patches (search-patches "gnutls-skip-trust-store-test.patch")) > > (sha256 > > (base32 > > - "0ghpyhhfa3nsraph6dws50jb3dc8g2cfl7dizdnyrm179fawakzp")))) > > + "0kayjxy3rr2y08jjimz5f0dx92pq3xjiaj2pdnsn15h1rp2k21pa")))) > > When I download the release from the official website[0], via the gnupg > ftp[1], I do get the file with the old hash[2]. > > I am not sure what is going on here, but the discrepancy should be > investigated and documented in the commit message. Just adjusting the > hash is probably not the best choice. > > I would assume the upstream[0] should be the authoritative source for > the hash, not the mirror. > > 0: https://gnutls.org/download.html > 1: https://www.gnupg.org/ftp/gcrypt/gnutls/v3.8/ > 2: 0ghpyhhfa3nsraph6dws50jb3dc8g2cfl7dizdnyrm179fawakzp > > > (build-system gnu-build-system) > > (arguments > > (list #:tests? (not (or (%current-target-system) > > Have a nice day, > Tomas > > -- > There are only two hard things in Computer Science: > cache invalidation, naming things and off-by-one errors.