At the time I'm writing this is a known issue in Tezos' codebase. Developers are working to fix this issue. It is fine setting export OPAMSOLVERTIMEOUT=300
as a temporary workaround, but beware it may not suffice; I'm regularly using OPAMSOLVERTIMEOUT=6000
. For a more robust workaround, apply the following patch (taken from this commit that isn't on master
):
diff --git a/src/tooling/tezos-tooling.opam b/src/tooling/tezos-tooling.opam
index e69888491..b0d9a8633 100644
--- a/src/tooling/tezos-tooling.opam
+++ b/src/tooling/tezos-tooling.opam
@@ -8,6 +8,11 @@ license: "MIT"
depends: [
"coq-of-ocaml" { = "2.2.1" }
"dune" { >= "2.0" }
+ "irmin" { = "2.2.0" }
+ "lwt" { = "5.3.0" }
+ "ocaml" { = "4.09.1" }
+ "ocaml-base-compiler" { = "4.09.1" }
+ "yojson" { = "1.7.0" }
"ocamlformat" { = "0.10" }
"bisect_ppx" { >= "2.3" }
]
that reduces the size of the constraints problem that opam's solver tackles.
[edit] Issue is now fixed, thanks to this commit: https://gitlab.com/tezos/tezos/-/commit/a07e391fd80a397be380de4b99b796aa530523ee Building a version after this commit should not cause a timeout anymore.
export OPAMSOLVERTIMEOUT=300