- 21 Mar, 2019 18 commits
-
-
xin authored
-
intrigeri authored
-
intrigeri authored
When preparing a bugfix release, or when one has correctly resolved merged conflicts in the previous step, this operation is a no-op. That's fine. But let's keep it around as a fail-safe, in case a mistake was made during the merge conflict resolution previously.
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
Most RMs should already have Git/SSH configuration to use torsocks for our canonical Git repo.
-
intrigeri authored
-
intrigeri authored
-
- 20 Mar, 2019 22 commits
-
-
IkiWiki authored
-
sajolida authored
-
sajolida authored
-
sajolida authored
-
IkiWiki authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
intrigeri authored
-
IkiWiki authored
-
sajolida authored
-
Ulrike Uhlig authored
-
Ulrike Uhlig authored
-
Ulrike Uhlig authored
-
Ulrike Uhlig authored
-
Ulrike Uhlig authored
-
intrigeri authored
-
intrigeri authored
-
Ulrike Uhlig authored
-
xin authored
-
intrigeri authored
When I automated this step, I was confused due to the problem I've fixed in the previous commit, so I wrote code that checks the published products against the IDF, while here we want to check the locally built products against the IDF. Granted, here again, this was valid *transitively*, because of the implicit (and so far correct) assumption that, at this point, we've already verified that what we have built matches SHA512SUMS.txt, and that the published products match SHA512SUMS.txt too, so both sets of products are equivalent at this point. Still, for the same reason as in the previous commit, let's simplify and directly do the check we want.
-
intrigeri authored
The previous instructions did the check we want… *transitively*: we checked that the published products matched SHA512SUMS.txt, in a step that claims to compare them with locally built products; this was only valid because of the implicit (and so far correct) assumption that, at this point, we've already verified that what you have built matches SHA512SUMS.txt. This adds complexity and makes it needlessly hard to understand and review this process. Instead, let's directly do the check we want.
-