At all 12 tinderbox images I experienced here sth like:
2021-06-02T00:24:04 >>> dev-util/ragel-7.0.4: 2 minutes, 12 seconds
2021-06-03T18:40:00 >>> dev-util/ragel-6.10: 55 seconds
2021-06-03T20:24:54 >>> dev-util/ragel-7.0.4: 2 minutes, 15 seconds
2021-06-03T21:52:21 >>> dev-util/ragel-6.10: 57 seconds
2021-06-03T22:06:18 >>> dev-util/ragel-7.0.4: 2 minutes, 10 seconds
2021-06-04T17:26:50 >>> dev-util/ragel-6.10: 1 minute, 15 seconds
2021-06-04T20:22:22 >>> dev-util/ragel-7.0.4: 2 minutes, 31 seconds
I'll attach all log files of emerge of that image here too.
The issue is generic and seems not to be related to special USE flag combos. All images are ~amd64.
Created attachment 713754 [details]
Created attachment 713757 [details]
Created attachment 713760 [details]
Created attachment 713763 [details]
Created attachment 713766 [details]
Created attachment 713769 [details]
Created attachment 713772 [details]
The problem is that some packages still declare dependencies on the older version, so we probably need to file bugs for them instead.
(In reply to Sam James from comment #8)
> The problem is that some packages still declare dependencies on the older
> version, so we probably need to file bugs for them instead.
My understanding is (at least for the ruby packages depending on ragel) that ragel 7 is not compatible with ragel 6 and required functionality is no longer present, so these dependencies won't be able to upgrade to ragel 7.
One thing to consider is to put ragel 6 and 7 in separate slots (with e.g. the ragel 6 slot providing a ragel-6 binary to run). I haven't looked at the feasibility of this. We would also need to consider desirability, especially if upstream no longer maintains ragel 6.
at least booth versions should block each other IMO if they can't coexists at the same system