view tests/test-module-imports.t @ 26457:7e81305092a0

demandimport: replace more references to _demandmod instances

_demandmod instances may be referenced by multiple importing modules.
Before this patch, the _demandmod instance only maintained a reference
to its first consumer when using the "from X import Y" syntax. This is
because we only created a single _demandmod instance (attached to the
parent X module). If multiple modules A and B performed
"from X import Y", we'd produce a single _demandmod instance
"demandmod" with the following references:

  X.Y = <demandmod>
  A.Y = <demandmod>
  B.Y = <demandmod>

The locals from the first consumer (A) would be stored in <demandmod1>.
When <demandmod1> was loaded, we'd look at the locals for the first
consumer and replace the symbol, if necessary. This resulted in state:

  X.Y = <module>
  A.Y = <module>
  B.Y = <demandmod>

B's reference to Y wasn't updated and was still using the proxy object
because we just didn't record that B had a reference to <demandmod> that
needed updating!

With this patch, we add support for tracking which modules in addition
to the initial importer have a reference to the _demandmod instance and
we replace those references at module load time.

In the case of posix.py, this fixes an issue where the "encoding" module
was being proxied, resulting in hundreds of thousands of
__getattribute__ lookups on the _demandmod instance during dirstate
operations on mozilla-central, speeding up execution by many
milliseconds. There are likely several other operation that benefit from
this change as well.

The new mechanism isn't perfect: references in locals (not globals) may
likely linger. So, if there is an import inside a function and a symbol
from that module is used in a hot loop, we could have unwanted overhead
from proxying through _demandmod. Non-global imports are discouraged
anyway. So hopefully this isn't a big deal in practice. We could
potentially deploy a code checker that bans use of attribute lookups of
function-level-imported modules inside loops.

This deficiency in theory could be avoided by storing the set of globals
and locals dicts to update in the _demandmod instance. However, I tried
this and it didn't work. One reason is that some globals are _demandmod
instances. We could work around this, but it's a bit more work. There
also might be other module import foo at play. The solution as
implemented is better than what we had and IMO is good enough for the
time being.

It's worth noting that this sub-optimal behavior was made worse by the
introduction of absolute_import and its recommended "from . import X"
syntax for importing modules from the "mercurial" package. If we ever
wrote performance tests, measuring the amount of module imports and
__getattribute__ proxy calls through _demandmod instances would be
something I'd have it check.

#require test-repo

  $ import_checker="$TESTDIR"/../contrib/import-checker.py

Run the doctests from the import checker, and make sure
it's working correctly.
  $ TERM=dumb
  $ export TERM
  $ python -m doctest $import_checker

Run additional tests for the import checker

  $ mkdir testpackage

  $ cat > testpackage/multiple.py << EOF
  > from __future__ import absolute_import
  > import os, sys
  > EOF

  $ cat > testpackage/unsorted.py << EOF
  > from __future__ import absolute_import
  > import sys
  > import os
  > EOF

  $ cat > testpackage/stdafterlocal.py << EOF
  > from __future__ import absolute_import
  > from . import unsorted
  > import os
  > EOF

  $ cat > testpackage/requirerelative.py << EOF
  > from __future__ import absolute_import
  > import testpackage.unsorted
  > EOF

  $ cat > testpackage/importalias.py << EOF
  > from __future__ import absolute_import
  > import ui
  > EOF

  $ cat > testpackage/relativestdlib.py << EOF
  > from __future__ import absolute_import
  > from .. import os
  > EOF

  $ cat > testpackage/symbolimport.py << EOF
  > from __future__ import absolute_import
  > from .unsorted import foo
  > EOF

  $ cat > testpackage/latesymbolimport.py << EOF
  > from __future__ import absolute_import
  > from . import unsorted
  > from mercurial.node import hex
  > EOF

  $ cat > testpackage/multiplegroups.py << EOF
  > from __future__ import absolute_import
  > from . import unsorted
  > from . import more
  > EOF

  $ mkdir testpackage/subpackage
  $ cat > testpackage/subpackage/levelpriority.py << EOF
  > from __future__ import absolute_import
  > from . import foo
  > from .. import parent
  > EOF

  $ cat > testpackage/sortedentries.py << EOF
  > from __future__ import absolute_import
  > from . import (
  >     foo,
  >     bar,
  > )
  > EOF

  $ cat > testpackage/importfromalias.py << EOF
  > from __future__ import absolute_import
  > from . import ui
  > EOF

  $ cat > testpackage/importfromrelative.py << EOF
  > from __future__ import absolute_import
  > from testpackage.unsorted import foo
  > EOF

  $ python "$import_checker" testpackage/*.py testpackage/subpackage/*.py
  testpackage/importalias.py ui module must be "as" aliased to uimod
  testpackage/importfromalias.py ui from testpackage must be "as" aliased to uimod
  testpackage/importfromrelative.py import should be relative: testpackage.unsorted
  testpackage/importfromrelative.py direct symbol import from testpackage.unsorted
  testpackage/latesymbolimport.py symbol import follows non-symbol import: mercurial.node
  testpackage/multiple.py multiple imported names: os, sys
  testpackage/multiplegroups.py multiple "from . import" statements
  testpackage/relativestdlib.py relative import of stdlib module
  testpackage/requirerelative.py import should be relative: testpackage.unsorted
  testpackage/sortedentries.py imports from testpackage not lexically sorted: bar < foo
  testpackage/stdafterlocal.py stdlib import follows local import: os
  testpackage/subpackage/levelpriority.py higher-level import should come first: testpackage
  testpackage/symbolimport.py direct symbol import from testpackage.unsorted
  testpackage/unsorted.py imports not lexically sorted: os < sys
  [1]

  $ cd "$TESTDIR"/..

There are a handful of cases here that require renaming a module so it
doesn't overlap with a stdlib module name. There are also some cycles
here that we should still endeavor to fix, and some cycles will be
hidden by deduplication algorithm in the cycle detector, so fixing
these may expose other cycles.

  $ hg locate 'mercurial/**.py' 'hgext/**.py' | sed 's-\\-/-g' | python "$import_checker" -
  Import cycle: hgext.largefiles.basestore -> hgext.largefiles.localstore -> hgext.largefiles.basestore
  [1]