view tests/test-confused-revert.t @ 25191:08d1ef09ed37

revset: optimize not public revset

This patvh speeds up the computation of the not public() changeset
and incidentally speed up the computation of divergents() changeset on our big
repo by 100x from 50% to 0.5% of the time spent in smartlog with evolve.

In this patch we optimize not public() to _notpublic() (new revset) and use
the work on phaseset (from the previous commit) to be able to compute
_notpublic() quickly.

We use a non-lazy approach making the assumption the number of notpublic
change will not be in the order of magnitude of the repo size. Adopting a
lazy approach gives a speedup of 5x (vs 100x) only due to the overhead of the
code for lazy generation.

  $ hg init
  $ echo foo > a
  $ hg add a
  $ hg commit -m "1"

  $ echo bar > b
  $ hg add b
  $ hg remove a

Should show a removed and b added:

  $ hg status
  A b
  R a

  $ hg revert --all
  undeleting a
  forgetting b

Should show b unknown and a back to normal:

  $ hg status
  ? b

  $ rm b

  $ hg co -C 0
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo foo-a > a
  $ hg commit -m "2a"

  $ hg co -C 0
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo foo-b > a
  $ hg commit -m "2b"
  created new head

  $ HGMERGE=true hg merge 1
  merging a
  0 files updated, 1 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

Should show foo-b:

  $ cat a
  foo-b

  $ echo bar > b
  $ hg add b
  $ rm a
  $ hg remove a

Should show a removed and b added:

  $ hg status
  A b
  R a

Revert should fail:

  $ hg revert
  abort: uncommitted merge with no revision specified
  (use "hg update" or see "hg help revert")
  [255]

Revert should be ok now:

  $ hg revert -r2 --all
  undeleting a
  forgetting b

Should show b unknown and a marked modified (merged):

  $ hg status
  M a
  ? b

Should show foo-b:

  $ cat a
  foo-b