view tests/test-patchbomb-tls.t @ 44118:f81c17ec303c

hgdemandimport: apply lazy module loading to sys.meta_path finders

Python's `sys.meta_path` finders are the primary objects whose job it
is to find a module at import time. When `import` is called, Python
iterates objects in this list and calls `o.find_spec(...)` to find
a `ModuleSpec` (or None if the module couldn't be found by that
finder). If no meta path finder can find a module, import fails.

One of the default meta path finders is `PathFinder`. Its job is to
import modules from the filesystem and is probably the most important
importer. This finder looks at `sys.path` and `sys.path_hooks` to do
its job.

The `ModuleSpec` returned by `MetaPathImporter.find_spec()` has a
`loader` attribute, which defines the concrete module loader to use.
`sys.path_hooks` is a hook point for teaching `PathFinder` to
instantiate custom loader types.

Previously, we injected a custom `sys.path_hook` that told `PathFinder`
to wrap the default loaders with a loader that creates a module object
that is lazy.

This approach worked. But its main limitation was that it only applied
to the `PathFinder` meta path importer. There are other meta path
importers that are registered. And in the case of PyOxidizer loading
modules from memory, `PathFinder` doesn't come into play since
PyOxidizer's own meta path importer was handling all imports.

This commit changes our approach to lazy module loading by proxying
all meta path importers. Specifically, we overload the `find_spec()`
method to swap in a wrapped loader on the `ModuleSpec` before it
is returned. The end result of this is all meta path importers should
be lazy.

As much as I would have loved to utilize .__class__ manipulation to
achieve this, some meta path importers are implemented in C/Rust
in such a way that they cannot be monkeypatched. This is why we
use __getattribute__ to define a proxy.

Also, this change could theoretically open us up to regressions in
meta path importers whose loader is creating module objects which
can't be monkeypatched. But I'm not aware of any of these in the
wild. So I think we'll be safe.

According to hyperfine, this change yields a decent startup time win of
5-6ms:

```
Benchmark #1: ~/.pyenv/versions/3.6.10/bin/python ./hg version
  Time (mean ± σ):      86.8 ms ±   0.5 ms    [User: 78.0 ms, System: 8.7 ms]
  Range (min … max):    86.0 ms …  89.1 ms    50 runs

  Time (mean ± σ):      81.1 ms ±   2.7 ms    [User: 74.5 ms, System: 6.5 ms]
  Range (min … max):    77.8 ms …  90.5 ms    50 runs

Benchmark #2: ~/.pyenv/versions/3.7.6/bin/python ./hg version
  Time (mean ± σ):      78.9 ms ±   0.6 ms    [User: 70.2 ms, System: 8.7 ms]
  Range (min … max):    78.1 ms …  81.2 ms    50 runs

  Time (mean ± σ):      73.4 ms ±   0.6 ms    [User: 65.3 ms, System: 8.0 ms]
  Range (min … max):    72.4 ms …  75.7 ms    50 runs

Benchmark #3: ~/.pyenv/versions/3.8.1/bin/python ./hg version
  Time (mean ± σ):      78.1 ms ±   0.6 ms    [User: 70.2 ms, System: 7.9 ms]
  Range (min … max):    77.4 ms …  80.9 ms    50 runs

  Time (mean ± σ):      72.1 ms ±   0.4 ms    [User: 64.4 ms, System: 7.6 ms]
  Range (min … max):    71.4 ms …  74.1 ms    50 runs
```

Differential Revision: https://phab.mercurial-scm.org/D7954

#require serve ssl

Set up SMTP server:

  $ CERTSDIR="$TESTDIR/sslcerts"
  $ cat "$CERTSDIR/priv.pem" "$CERTSDIR/pub.pem" >> server.pem

  $ "$PYTHON" "$TESTDIR/dummysmtpd.py" -p $HGPORT --pid-file a.pid -d \
  > --tls smtps --certificate `pwd`/server.pem
  listening at localhost:$HGPORT (?)
  $ cat a.pid >> $DAEMON_PIDS

Set up repository:

  $ hg init t
  $ cd t
  $ cat <<EOF >> .hg/hgrc
  > [extensions]
  > patchbomb =
  > [email]
  > method = smtp
  > [smtp]
  > host = localhost
  > port = $HGPORT
  > tls = smtps
  > EOF

  $ echo a > a
  $ hg commit -Ama -d '1 0'
  adding a

Utility functions:

  $ DISABLECACERTS=
  $ try () {
  >   hg email $DISABLECACERTS -f quux -t foo -c bar -r tip "$@"
  > }

Our test cert is not signed by a trusted CA. It should fail to verify if
we are able to load CA certs:

#if sslcontext defaultcacerts no-defaultcacertsloaded
  $ try
  this patch series consists of 1 patches.
  
  
  (an attempt was made to load CA certificates but none were loaded; see https://mercurial-scm.org/wiki/SecureConnections for how to configure Mercurial to avoid this error)
  (?i)abort: .*?certificate.verify.failed.* (re)
  [255]
#endif

#if no-sslcontext defaultcacerts
  $ try
  this patch series consists of 1 patches.
  
  
  warning: connecting to localhost using legacy security technology (TLS 1.0); see https://mercurial-scm.org/wiki/SecureConnections for more info
  (using CA certificates from *; if you see this message, your Mercurial install is not properly configured; see https://mercurial-scm.org/wiki/SecureConnections for how to configure Mercurial to avoid this message) (glob) (?)
  (?i)abort: .*?certificate.verify.failed.* (re)
  [255]
#endif

#if defaultcacertsloaded
  $ try
  this patch series consists of 1 patches.
  
  
  warning: connecting to localhost using legacy security technology (TLS 1.0); see https://mercurial-scm.org/wiki/SecureConnections for more info (?)
  (using CA certificates from *; if you see this message, your Mercurial install is not properly configured; see https://mercurial-scm.org/wiki/SecureConnections for how to configure Mercurial to avoid this message) (glob) (?)
  (the full certificate chain may not be available locally; see "hg help debugssl") (windows !)
  (?i)abort: .*?certificate.verify.failed.* (re)
  [255]

#endif

#if no-defaultcacerts
  $ try
  this patch series consists of 1 patches.
  
  
  (unable to load * certificates; see https://mercurial-scm.org/wiki/SecureConnections for how to configure Mercurial to avoid this message) (glob) (?)
  abort: localhost certificate error: no certificate received
  (set hostsecurity.localhost:certfingerprints=sha256:62:09:97:2f:97:60:e3:65:8f:12:5d:78:9e:35:a1:36:7a:65:4b:0e:9f:ac:db:c3:bc:6e:b6:a3:c0:16:e0:30 config setting or use --insecure to connect insecurely)
  [255]
#endif

  $ DISABLECACERTS="--config devel.disableloaddefaultcerts=true"

Without certificates:

  $ try --debug
  this patch series consists of 1 patches.
  
  
  (using smtps)
  sending mail: smtp host localhost, port * (glob)
  warning: connecting to localhost using legacy security technology (TLS 1.0); see https://mercurial-scm.org/wiki/SecureConnections for more info (?)
  (verifying remote certificate)
  abort: unable to verify security of localhost (no loaded CA certificates); refusing to connect
  (see https://mercurial-scm.org/wiki/SecureConnections for how to configure Mercurial to avoid this error or set hostsecurity.localhost:fingerprints=sha256:20:de:b3:ad:b4:cd:a5:42:f0:74:41:1c:a2:70:1e:da:6e:c0:5c:16:9e:e7:22:0f:f1:b7:e5:6e:e4:92:af:7e to trust this server)
  [255]

With global certificates:

  $ try --debug --config web.cacerts="$CERTSDIR/pub.pem"
  this patch series consists of 1 patches.
  
  
  (using smtps)
  sending mail: smtp host localhost, port * (glob)
  warning: connecting to localhost using legacy security technology (TLS 1.0); see https://mercurial-scm.org/wiki/SecureConnections for more info (?)
  (verifying remote certificate)
  sending [PATCH] a ...

With invalid certificates:

  $ try --config web.cacerts="$CERTSDIR/pub-other.pem"
  this patch series consists of 1 patches.
  
  
  warning: connecting to localhost using legacy security technology (TLS 1.0); see https://mercurial-scm.org/wiki/SecureConnections for more info (?)
  (the full certificate chain may not be available locally; see "hg help debugssl") (windows !)
  (?i)abort: .*?certificate.verify.failed.* (re)
  [255]

  $ cd ..