view tests/test-evolve-split.t @ 6791:2ef4e26c4bc0 stable

topic: add more branchcache entries to test-topic-flow-publish-flag.t

It seems these new entries started appearing after the recent-ish refactoring
of branchcache in core. Now the server and the client have the same branch
cache data on disk, which is probably for the best.

The actual output with changeset hashes is:
branch2-base:c0155e8ec55a9b46d415d5b13a319d0c80e26e9d o double//slash
branch2-immutable:c0155e8ec55a9b46d415d5b13a319d0c80e26e9d o double//slash
branch2-served:256a2cb64fd8ada4ff18d57606d212f094cb50fb o double//slash

c0155e8ec55a is public, so it's correct to be in "base" and "immutable" repo
views. 256a2cb64fd8 is draft (without topic), so it's expected to be in the
"served" view.

Check that evolve can handle split commits
------------------------------------------

  $ cat >> $HGRCPATH <<EOF
  > [ui]
  > logtemplate = {rev}:{node|short}@{branch}({phase}) {desc|firstline}\n
  > [extensions]
  > EOF
  $ echo "evolve=$(echo $(dirname $TESTDIR))/hgext3rd/evolve/" >> $HGRCPATH
  $ mkcommit() {
  >    echo "$1" > "$1"
  >    hg add "$1"
  >    hg ci -m "add $1"
  > }

  $ hg init split
  $ cd split
  $ mkcommit aa

Create a split commit
  $ printf "oo" > oo;
  $ printf "pp" > pp;
  $ hg add oo pp
  $ hg commit -m "oo+pp"
  $ mkcommit uu
  $ hg up 0
  0 files updated, 0 files merged, 3 files removed, 0 files unresolved
  $ printf "oo" > oo;
  $ hg add oo
  $ hg commit -m "_oo"
  created new head
  $ printf "pp" > pp;
  $ hg add pp
  $ hg commit -m "_pp"
  $ hg prune --successor "desc(_oo) + desc(_pp)" -r "desc('oo+pp')" --split
  1 changesets pruned
  1 new orphan changesets
  $ hg log -G
  @  4:d0dcf24cddd3@default(draft) _pp
  |
  o  3:a7fdfda64c08@default(draft) _oo
  |
  | *  2:f52200b086ca@default(draft) add uu
  | |
  | x  1:d55647aaa0c6@default(draft) oo+pp
  |/
  o  0:58663bb03074@default(draft) add aa
  
  $ hg evolve --rev "0::"
  move:[2] add uu
  atop:[4] _pp

  $ cd ..
  $ hg init split-merged
  $ cd split-merged
  $ mkcommit aa

Split the changeset such that the successors don't have a single root and there's an unrelated changeset in between
  $ printf "oo" > oo;
  $ printf "pp" > pp;
  $ printf "qq" > qq;
  $ hg add oo pp qq
  $ hg commit -m "oo+pp+qq"
  $ mkcommit uu
  $ hg up 0
  0 files updated, 0 files merged, 4 files removed, 0 files unresolved
  $ printf "oo" > oo;
  $ hg add oo
  $ hg commit -m "_oo"
  created new head
  $ hg up 0
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ printf "pp" > pp;
  $ hg add pp
  $ hg commit -m "_pp"
  created new head
  $ hg merge 3
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg ci -m 'merge oo and pp'
  $ printf "qq" > qq;
  $ hg add qq
  $ hg commit -m "_qq"
  $ hg prune --successor "desc(_oo) + desc(_pp) + desc(_qq)" -r "desc('oo+pp+qq')" --split
  1 changesets pruned
  1 new orphan changesets
  $ hg log -G
  @  6:ea5b1e180c04@default(draft) _qq
  |
  o    5:bf7c32161b4b@default(draft) merge oo and pp
  |\
  | o  4:ece0aaa22eb7@default(draft) _pp
  | |
  o |  3:a7fdfda64c08@default(draft) _oo
  |/
  | *  2:cc56c47d84b3@default(draft) add uu
  | |
  | x  1:575a7380a87d@default(draft) oo+pp+qq
  |/
  o  0:58663bb03074@default(draft) add aa
  
  $ hg evolve --rev "0::"
  move:[2] add uu
  atop:[6] _qq