view tests/test-audit-subrepo.t @ 51681:522b4d729e89

mmap: populate the mapping by default

Without pre-population, accessing all data through a mmap can result in many
pagefault, reducing performance significantly. If the mmap is prepopulated, the
performance can no longer get slower than a full read.
(See benchmark number below)

In some cases were very few data is read, prepopulating can be overkill and
slower than populating on access (through page fault). So that behavior can be
controlled when the caller can pre-determine the best behavior.
(See benchmark number below)

In addition, testing with populating in a secondary thread yield great result
combining the best of each approach. This might be implemented in later
changesets.

In all cases, using mmap has a great effect on memory usage when many processes
run in parallel on the same machine.

### Benchmarks

# What did I run

A couple of month back I ran a large benchmark campaign to assess the impact of
various approach for using mmap with the revlog (and other files), it
highlighted a few benchmarks that capture the impact of the changes well. So to
validate this change I checked the following:

- log command displaying various revisions
  (read the changelog index)
- log command displaying the patch of listed revisions
  (read the changelog index, the manifest index and a few files indexes)
- unbundling a few revisions
  (read and write changelog, manifest and few files indexes, and walk the graph
  to update some cache)
- pushing a few revisions
  (read and write changelog, manifest and few files indexes, walk the graph to
  update some cache, performs various accesses locally and remotely during
  discovery)

Benchmarks were run using the default module policy (c+py) and the rust one. No
significant difference were found between the two implementation, so we will
present result using the default policy (unless otherwise specified).

I ran them on a few repositories :

- mercurial: a "public changeset only" copy of mercurial from 2018-08-01 using
  zstd compression and sparse-revlog

- pypy: a copy of pypy from 2018-08-01 using zstd compression and sparse-revlog

- netbeans: a copy of netbeans from 2018-08-01 using zstd compression and
  sparse-revlog

- mozilla-try: a copy of mozilla-try from 2019-02-18 using zstd compression and
  sparse-revlog

- mozilla-try persistent-nodemap: Same as the above but with a persistent
  nodemap. Used for the log --patch benchmark only

# Results

For the smaller repositories (mercurial, pypy), the impact of mmap is almost
imperceptible, other cost dominating the operation. The impact of prepopulating
is undiscernible in the benchmark we ran.

For larger repositories the benchmark support explanation given above:

On netbeans, the log can be about 1% faster without repopulation (for a
difference < 100ms) but unbundle becomes a bit slower, even when small.

### data-env-vars.name                             = netbeans-2018-08-01-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
  # benchmark.variants.source                      = unbundle
  # benchmark.variants.verbosity                   = quiet
with-populate: 0.240157
no-populate:   0.265087  (+10.38%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)
 ## benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 0.771919
no-populate:   0.792025   (+2.60%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)


For mozilla-try, the "slow down" from pre-populate for small `hg log` is more
visible, but still small in absolute time. (using rust value for the persistent
nodemap value to be relevant).

### data-env-vars.name            = mozilla-try-2019-02-18-ds2-pnm
  # benchmark.name                = hg.command.log
  # bin-env-vars.hg.flavor        = rust
  # benchmark.variants.patch      = yes
  # benchmark.variants.limit-rev  = 1
with-populate: 0.237813
no-populate:   0.229452   (-3.52%, -0.01)
  # benchmark.variants.limit-rev  = 10
  # benchmark.variants.patch      = yes
with-populate: 1.213578
no-populate:   1.205189
### data-env-vars.name            = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.variants.limit-rev  = 1000
  # benchmark.variants.patch      = no
  # benchmark.variants.rev        = tip
with-populate: 0.198607
no-populate:   0.195038   (-1.80%, -0.00)

However pre-populating provide a significant boost on more complex operations
like unbundle or push:

### data-env-vars.name                             = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 4.798632
no-populate:   4.953295   (+3.22%, +0.15)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 4.903618
no-populate:   5.014963   (+2.27%, +0.11)
 ## benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 1.423411
no-populate:   1.585365  (+11.38%, +0.16)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.537909
no-populate:   1.688489   (+9.79%, +0.15)

Test illegal name
-----------------

on commit:

  $ hg init hgname
  $ cd hgname
  $ mkdir sub
  $ hg init sub/.hg
  $ echo 'sub/.hg = sub/.hg' >> .hgsub
  $ hg ci -qAm 'add subrepo "sub/.hg"'
  abort: path 'sub/.hg' is inside nested repo 'sub'
  [10]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "sub/.hg"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +sub/.hg = sub/.hg
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 sub/.hg
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q hgname hgname2
  abort: path 'sub/.hg' is inside nested repo 'sub'
  [10]

Test absolute path
------------------

on commit:

  $ hg init absolutepath
  $ cd absolutepath
  $ hg init sub
  $ echo '/sub = sub' >> .hgsub
  $ hg ci -qAm 'add subrepo "/sub"'
  abort: path contains illegal component: /sub
  [10]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "/sub"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +/sub = sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 /sub
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q absolutepath absolutepath2
  abort: path contains illegal component: /sub
  [10]

Test root path
--------------

on commit:

  $ hg init rootpath
  $ cd rootpath
  $ hg init sub
  $ echo '/ = sub' >> .hgsub
  $ hg ci -qAm 'add subrepo "/"'
  abort: path ends in directory separator: /
  [10]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "/"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +/ = sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 /
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q rootpath rootpath2
  abort: path ends in directory separator: /
  [10]

Test empty path
---------------

on commit:

  $ hg init emptypath
  $ cd emptypath
  $ hg init sub
  $ echo '= sub' >> .hgsub
  $ hg ci -qAm 'add subrepo ""'
  config error at .hgsub:1: = sub
  [30]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo ""' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > += sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q emptypath emptypath2
  config error at .hgsub:1: = sub
  [30]

Test current path
-----------------

on commit:

  $ hg init currentpath
  $ cd currentpath
  $ hg init sub
  $ echo '. = sub' >> .hgsub
  $ hg ci -qAm 'add subrepo "."'
  abort: subrepo path contains illegal component: .
  [255]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "."' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +.= sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 .
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q currentpath currentpath2
  abort: subrepo path contains illegal component: .
  [255]

Test outer path
---------------

on commit:

  $ mkdir outerpath
  $ cd outerpath
  $ hg init main
  $ cd main
  $ hg init ../sub
  $ echo '../sub = ../sub' >> .hgsub
  $ hg ci -qAm 'add subrepo "../sub"'
  abort: path contains illegal component: ../sub
  [10]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "../sub"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +../sub = ../sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 ../sub
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q main main2
  abort: path contains illegal component: ../sub
  [10]
  $ cd ..

Test variable expansion
-----------------------

Subrepository paths shouldn't be expanded, but we fail to handle them
properly. Any local repository paths are expanded.

on commit:

  $ mkdir envvar
  $ cd envvar
  $ hg init main
  $ cd main
  $ hg init sub1
  $ cat <<'EOF' > sub1/hgrc
  > [hooks]
  > log = echo pwned
  > EOF
  $ hg -R sub1 ci -qAm 'add sub1 files'
  $ hg -R sub1 log -r. -T '{node}\n'
  39eb4b4d3e096527668784893a9280578a8f38b8
  $ echo '$SUB = sub1' >> .hgsub
  $ SUB=sub1 hg ci -qAm 'add subrepo "$SUB"'
  abort: subrepo path contains illegal component: $SUB
  [255]

prepare tampered repo (including the changes above as two commits):

  $ hg import --bypass -qm 'add subrepo "$SUB"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +$SUB = sub1
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 $SUB
  > EOF
  $ hg debugsetparents 0
  $ hg import --bypass -qm 'update subrepo "$SUB"' - <<'EOF'
  > diff --git a/.hgsubstate b/.hgsubstate
  > --- a/.hgsubstate
  > +++ b/.hgsubstate
  > @@ -1,1 +1,1 @@
  > -0000000000000000000000000000000000000000 $SUB
  > +39eb4b4d3e096527668784893a9280578a8f38b8 $SUB
  > EOF
  $ cd ..

on clone (and update) with various substitutions:

  $ hg clone -q main main2
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main2
  .hg
  .hgsub
  .hgsubstate

  $ SUB=sub1 hg clone -q main main3
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main3
  .hg
  .hgsub
  .hgsubstate

  $ SUB=sub2 hg clone -q main main4
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main4
  .hg
  .hgsub
  .hgsubstate

on clone empty subrepo into .hg, then pull (and update), which at least fails:

  $ SUB=.hg hg clone -qr0 main main5
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main5
  .hg
  .hgsub
  .hgsubstate
  $ test -d main5/.hg/.hg
  [1]
  $ SUB=.hg hg -R main5 pull -u
  pulling from $TESTTMP/envvar/main
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files
  new changesets 7a2f0e59146f
  .hgsubstate: untracked file differs
  abort: untracked files in working directory differ from files in requested revision
  [20]
  $ cat main5/.hg/hgrc | grep pwned
  [1]

on clone (and update) into .hg, which at least fails:

  $ SUB=.hg hg clone -q main main6
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main6
  .hg
  .hgsub
  .hgsubstate
  $ cat main6/.hg/hgrc | grep pwned
  [1]

on clone (and update) into .hg/* subdir:

  $ SUB=.hg/foo hg clone -q main main7
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main7
  .hg
  .hgsub
  .hgsubstate
  $ test -d main7/.hg/.hg
  [1]

on clone (and update) into outer tree:

  $ SUB=../out-of-tree-write hg clone -q main main8
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main8
  .hg
  .hgsub
  .hgsubstate

on clone (and update) into e.g. $HOME, which doesn't work since subrepo paths
are concatenated prior to variable expansion:

  $ SUB="$TESTTMP/envvar/fakehome" hg clone -q main main9
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A main9 | wc -l
  \s*3 (re)

  $ ls
  main
  main2
  main3
  main4
  main5
  main6
  main7
  main8
  main9
  $ cd ..

Test tilde
----------

The leading tilde may be expanded to $HOME, but it can be a valid subrepo
path in theory. However, we want to prohibit it as there might be unsafe
handling of such paths.

on commit:

  $ hg init tilde
  $ cd tilde
  $ hg init './~'
  $ echo '~ = ~' >> .hgsub
  $ hg ci -qAm 'add subrepo "~"'
  abort: subrepo path contains illegal component: ~
  [255]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "~"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +~ = ~
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 ~
  > EOF
  $ cd ..

on clone (and update):

  $ hg clone -q tilde tilde2
  abort: subrepo path contains illegal component: ~
  [255]

Test direct symlink traversal
-----------------------------

#if symlink

on commit:

  $ mkdir hgsymdir
  $ hg init hgsymdir/root
  $ cd hgsymdir/root
  $ ln -s ../out
  $ hg ci -qAm 'add symlink "out"'
  $ hg init ../out
  $ echo 'out = out' >> .hgsub
  $ hg ci -qAm 'add subrepo "out"'
  abort: subrepo 'out' traverses symbolic link
  [255]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "out"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +out = out
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 out
  > EOF
  $ cd ../..

on clone (and update):

  $ mkdir hgsymdir2
  $ hg clone -q hgsymdir/root hgsymdir2/root
  abort: subrepo 'out' traverses symbolic link
  [255]
  $ ls hgsymdir2
  root

#endif

Test indirect symlink traversal
-------------------------------

#if symlink

on commit:

  $ mkdir hgsymin
  $ hg init hgsymin/root
  $ cd hgsymin/root
  $ ln -s ../out
  $ hg ci -qAm 'add symlink "out"'
  $ mkdir ../out
  $ hg init ../out/sub
  $ echo 'out/sub = out/sub' >> .hgsub
  $ hg ci -qAm 'add subrepo "out/sub"'
  abort: path 'out/sub' traverses symbolic link 'out'
  [255]

prepare tampered repo (including the commit above):

  $ hg import --bypass -qm 'add subrepo "out/sub"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +out/sub = out/sub
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 out/sub
  > EOF
  $ cd ../..

on clone (and update):

  $ mkdir hgsymin2
  $ hg clone -q hgsymin/root hgsymin2/root
  abort: path 'out/sub' traverses symbolic link 'out'
  [255]
  $ ls hgsymin2
  root

#endif

Test symlink traversal by variable expansion
--------------------------------------------

#if symlink

  $ FAKEHOME="$TESTTMP/envvarsym/fakehome"

on commit:

  $ mkdir envvarsym
  $ cd envvarsym
  $ hg init main
  $ cd main
  $ ln -s "`echo "$FAKEHOME" | sed 's|\(.\)/.*|\1|'`"
  $ hg ci -qAm 'add symlink to top-level system directory'

  $ hg init sub1
  $ echo pwned > sub1/pwned
  $ hg -R sub1 ci -qAm 'add sub1 files'
  $ hg -R sub1 log -r. -T '{node}\n'
  f40c9134ba1b6961e12f250868823f0092fb68a8
  $ echo '$SUB = sub1' >> .hgsub
  $ SUB="$FAKEHOME" hg ci -qAm 'add subrepo "$SUB"'
  abort: subrepo path contains illegal component: $SUB
  [255]

prepare tampered repo (including the changes above as two commits):

  $ hg import --bypass -qm 'add subrepo "$SUB"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +$SUB = sub1
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 $SUB
  > EOF
  $ hg debugsetparents 1
  $ hg import --bypass -qm 'update subrepo "$SUB"' - <<'EOF'
  > diff --git a/.hgsubstate b/.hgsubstate
  > --- a/.hgsubstate
  > +++ b/.hgsubstate
  > @@ -1,1 +1,1 @@
  > -0000000000000000000000000000000000000000 $SUB
  > +f40c9134ba1b6961e12f250868823f0092fb68a8 $SUB
  > EOF
  $ cd ..

on clone (and update) without fakehome directory:

  $ rm -fR "$FAKEHOME"
  $ SUB="$FAKEHOME" hg clone -q main main2
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ test -d "$FAKEHOME"
  [1]

on clone (and update) with empty fakehome directory:

  $ rm -fR "$FAKEHOME"
  $ mkdir "$FAKEHOME"
  $ SUB="$FAKEHOME" hg clone -q main main3
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls "$FAKEHOME"

on clone (and update) with non-empty fakehome directory:

  $ rm -fR "$FAKEHOME"
  $ mkdir "$FAKEHOME"
  $ touch "$FAKEHOME/a"
  $ SUB="$FAKEHOME" hg clone -q main main4
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls "$FAKEHOME"
  a

on clone empty subrepo with non-empty fakehome directory,
then pull (and update):

  $ rm -fR "$FAKEHOME"
  $ mkdir "$FAKEHOME"
  $ touch "$FAKEHOME/a"
  $ SUB="$FAKEHOME" hg clone -qr1 main main5
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls "$FAKEHOME"
  a
  $ test -d "$FAKEHOME/.hg"
  [1]
  $ SUB="$FAKEHOME" hg -R main5 pull -u
  pulling from $TESTTMP/envvarsym/main
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files
  new changesets * (glob)
  .hgsubstate: untracked file differs
  abort: untracked files in working directory differ from files in requested revision
  [20]
  $ ls "$FAKEHOME"
  a
  $ test -d "$FAKEHOME/.hg"
  [1]

on clone empty subrepo with hg-managed fakehome directory,
then pull (and update):

  $ rm -fR "$FAKEHOME"
  $ hg init "$FAKEHOME"
  $ touch "$FAKEHOME/a"
  $ hg -R "$FAKEHOME" ci -qAm 'add fakehome file'
  $ SUB="$FAKEHOME" hg clone -qr1 main main6
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A "$FAKEHOME"
  .hg
  a
  $ SUB="$FAKEHOME" hg -R main6 pull -u
  pulling from $TESTTMP/envvarsym/main
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files
  new changesets * (glob)
  .hgsubstate: untracked file differs
  abort: untracked files in working directory differ from files in requested revision
  [20]
  $ ls -A "$FAKEHOME"
  .hg
  a

on clone only symlink with hg-managed fakehome directory,
then pull (and update):

  $ rm -fR "$FAKEHOME"
  $ hg init "$FAKEHOME"
  $ touch "$FAKEHOME/a"
  $ hg -R "$FAKEHOME" ci -qAm 'add fakehome file'
  $ SUB="$FAKEHOME" hg clone -qr0 main main7
  $ ls -A "$FAKEHOME"
  .hg
  a
  $ SUB="$FAKEHOME" hg -R main7 pull -uf
  pulling from $TESTTMP/envvarsym/main
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 3 changes to 2 files
  new changesets * (glob)
  abort: subrepo path contains illegal component: $SUB
  [255]
  $ ls -A "$FAKEHOME"
  .hg
  a

  $ cd ..

#endif

Test drive letter
-----------------

Windows has a weird relative path that can change the drive letter, which
should also be prohibited on Windows.

prepare tampered repo:

  $ hg init driveletter
  $ cd driveletter
  $ hg import --bypass -qm 'add subrepo "X:"' - <<'EOF'
  > diff --git a/.hgsub b/.hgsub
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsub
  > @@ -0,0 +1,1 @@
  > +X: = foo
  > diff --git a/.hgsubstate b/.hgsubstate
  > new file mode 100644
  > --- /dev/null
  > +++ b/.hgsubstate
  > @@ -0,0 +1,1 @@
  > +0000000000000000000000000000000000000000 X:
  > EOF
  $ cd ..

on clone (and update):

#if windows

  $ hg clone -q driveletter driveletter2
  abort: path contains illegal component: X:
  [10]

#else

  $ hg clone -q driveletter driveletter2
  $ ls -A driveletter2
  .hg
  .hgsub
  .hgsubstate
  X:

#endif