couchbase-lite-core icon indicating copy to clipboard operation
couchbase-lite-core copied to clipboard

CBL-5947: Merge Beryllium into master

Open callumbirks opened this issue 1 year ago • 5 comments

The merge is not as bad as I thought it would be. A few conflicts needed resolving, but no manual adjustment needed to take place. I've run the VectorSearch tests locally and they are passing.

callumbirks avatar Jul 01 '24 10:07 callumbirks

Code Coverage Results:

Type Percentage
branches 67.15
functions 79.0
instantiations 35.1
lines 78.66
regions 74.49

cbl-bot avatar Jul 01 '24 11:07 cbl-bot

Blackduck failure is a strange one - seems to have trouble with some mysterious non-ascii character in the git log output. I ran the same git log command locally and used grep, and then manual looking, and could not find this mysterious character, so I think we can ignore it.

callumbirks avatar Jul 03 '24 12:07 callumbirks

Blocked by https://github.com/couchbase/couchbase-lite-core-EE/pull/43

callumbirks avatar Jul 03 '24 12:07 callumbirks

Reverted "Give CMake build -Werror" because it causes many errors in linux build. As its not a simple 4-line change, going to create a separate ticket for that.

callumbirks avatar Jul 03 '24 16:07 callumbirks

@jianminzhao If you could give this a review now, the only failure is a strange error from the blackduck manifest check which seems to have a problem with some non-ASCII character in the git log output (which I couldn't find)

callumbirks avatar Jul 04 '24 08:07 callumbirks

@jianminzhao If you could give this a review now, the only failure is a strange error from the blackduck manifest check which seems to have a problem with some non-ASCII character in the git log output (which I couldn't find)

Change ascii to utf-8 and it should probably be fine

borrrden avatar Jul 09 '24 00:07 borrrden

Now the big question: Which parts of this were conflicts that need to be reviewed and which parts applied cleanly?

borrrden avatar Jul 11 '24 00:07 borrrden

Now the big question: Which parts of this were conflicts that need to be reviewed and which parts applied cleanly?

@borrrden Well, there were roughly 12 files with conflicts IIRC, each with about 3 conflicts on average. Almost all of those conflicts were ones where I just needed to pick between the two, because they're different versions of the same thing. So this meant mostly choosing the version from 3.2.

If we think it needs it, I could go through the conflicts again in a meeting?

callumbirks avatar Jul 11 '24 04:07 callumbirks

Excellent, so no conflicts that were too complicated? That sounds good enough to me

borrrden avatar Jul 12 '24 22:07 borrrden