Mercurial > hg-stable
view tests/test-unified-test.t @ 32200:4d504e541d3d
rebase: use matcher to optimize manifestmerge
The old merge code would call manifestmerge and calculate the complete diff
between the source to the destination. In many cases, like rebase, the vast
majority of differences between the source and destination are irrelevant
because they are differences between the destination and the common ancestor
only, and therefore don't affect the merge. Since most actions are 'keep', all
the effort to compute them is wasted.
Instead, let's compute the difference between the source and the common ancestor
and only perform the diff of those files against the merge destination. When
using treemanifest, this lets us avoid loading almost the entire tree when
rebasing from a very old ancestor. This speeds up rebase of an old stack of 27
commits by 20x.
In mozilla-central, without treemanifest, when rebasing a commit from
default~100000 to default, this speeds up the manifestmerge step from 2.6s to
1.2s. However, the additional diff adds an overhead to all manifestmerge calls,
especially for flat manifests. When rebasing a commit from default~1 to default
it appears to add 100ms in mozilla-central. While we could put this optimization
behind a flag, I think the fact that it makes merge O(number of changes being
applied) instead of O(number of changes between X and Y) justifies it.
author | Durham Goode <durham@fb.com> |
---|---|
date | Wed, 03 May 2017 10:43:59 -0700 |
parents | 6a98f9408a50 |
children | 4441705b7111 |
line wrap: on
line source
Test that the syntax of "unified tests" is properly processed ============================================================== Simple commands: $ echo foo foo $ printf 'oh no' oh no (no-eol) $ printf 'bar\nbaz\n' | cat bar baz Multi-line command: $ foo() { > echo bar > } $ foo bar Return codes before inline python: $ sh -c 'exit 1' [1] Doctest commands: >>> from __future__ import print_function >>> print('foo') foo $ echo interleaved interleaved >>> for c in 'xyz': ... print(c) x y z >>> print() >>> foo = 'global name' >>> def func(): ... print(foo, 'should be visible in func()') >>> func() global name should be visible in func() >>> print('''multiline ... string''') multiline string Regular expressions: $ echo foobarbaz foobar.* (re) $ echo barbazquux .*quux.* (re) Globs: $ printf '* \\foobarbaz {10}\n' \* \\fo?bar* {10} (glob) Literal match ending in " (re)": $ echo 'foo (re)' foo (re) Windows: \r\n is handled like \n and can be escaped: #if windows $ printf 'crlf\r\ncr\r\tcrlf\r\ncrlf\r\n' crlf cr\r (no-eol) (esc) \tcrlf (esc) crlf\r (esc) #endif Combining esc with other markups - and handling lines ending with \r instead of \n: $ printf 'foo/bar\r' fo?/bar\r (no-eol) (glob) (esc) #if windows $ printf 'foo\\bar\r' foo/bar\r (no-eol) (glob) (esc) #endif $ printf 'foo/bar\rfoo/bar\r' foo.bar\r \(no-eol\) (re) (esc) foo.bar\r \(no-eol\) (re) testing hghave $ hghave true $ hghave false skipped: missing feature: nail clipper [1] $ hghave no-true skipped: system supports yak shaving [1] $ hghave no-false Conditional sections based on hghave: #if true $ echo tested tested #else $ echo skipped #endif #if false $ echo skipped #else $ echo tested tested #endif #if no-false $ echo tested tested #else $ echo skipped #endif #if no-true $ echo skipped #else $ echo tested tested #endif Exit code: $ (exit 1) [1]