stest: If one of our secnets dies, call the test a failure

This involves use of TclX's `signal' facility.  In my tests it was
easy to make Tcl deadlock by doing too much work in the signal
handler.  In particular reaping children is a bad idea.  Also signals
are not blocked during the signal handler so it would have to be
reentrant.

Instead, use `after idle'.  That is quite soon enough for the reap to
run, and in my tests with TclX 8.4 it all works properly.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Use `finish' for success exits

vwait is a bit funny.  This is better, and it also kills our child
secnets.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Kill our child secnets when we call finish

This stops them hanging around and, often, printing more stuff after
the test has actually failed.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Record the pids of of the secnets we spawn

Nothing uses this yet.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Replace the call to `exit 1' with a new proc `finish'

This is going to be used for success exits too, in a moment.
This will let us do more work when we are exiting.

The only change for now is some extra stderr output.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Support not running secnet ourselves

Now the user can set SECNET_STEST_DIVERT_inside=1 (say) and instead of
actually running secnet, we print the rune we would run and ask the
user when they have started it.

A freshly started secnet (with this config, at least) just sits there
waiting for its first packet to deal with.  So if the user does this
for the 2nd secnet, that's fine - the first one will happily wait
until the test actually gets going.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Don't prefix ./ with another ./, in socktmp

Makes things slightly prettier.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Print the special env vars we pass to secnet

We don't dump the whole environment, just selected bits.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Print the secnet runes we run

We can't use prexec here because we aren't using exec at all.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

tests, mtest: Provide and use `prexec' for helping with debugging

This prints the make-secnet-sites runes we are about to run.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dir-locals: Provide python-indent-offset too

python-indent is obsolete since Emacs 24.3.  It has been made into an
alias for python-indent-offset, but the alias is not marked safe as a
file variable.  Specifying python-indent-offset should make this work
on all relevant versions of Emacs (at least in the default
configuration).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Add debug logs

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Use new `prefix' option.

This lets us distinguish output from our two secnets.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Use stderr, not tty, for logging

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

logfile: New `prefix' option.

This allows the config file to add a fixed string to log messages.
This will be useful in our test suite, which mixes up output from
two instances of secnet.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

logfile: Do not log differently with --nodetach

This is actually quite anoying.  There is no good reason for leaving
off the timestamps in tests, or manual debugging, etc.

If at some point we want to support running under some kind of
log-capturing daemon supervisor which timestamps everything, we can
easily make this conditional again.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

logfile: Remove redundant check for st->f

st->f is now always non-null.  If no filename is specified, it is
stderr.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

logfile: Log to stderr by default (`filename' key now optional)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

secnet: Make stderr line buffered

This helps not interleave output in tests.  We write output in lines,
so this is fine.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test-example: Fix a dependency on Subdir.mk

This should be *this* directory's Subdir.mk.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test-example: Replace a literal with $@

No functional change.  This makes the command a bit more general.
At some much later point this will make a future commit simpler...

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

comprehensive-test: Test that clean target works

We have to have a list of exceptions for the output of configure.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Add some missing clean and cdeps hooks

Tested with
  git-clean -nX | grep -v \*~ | grep -v '~$' | grep -v Makefile | grep -v '\.mk$'

We now clean, additionally,
  .msgcode-test.d
  .version.d
  ipaddrset.confirm

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Switch to using subdirmk's clean and cdeps

Even in the toplevel.  This cleans some more things but doesn't fix
all the lack of clean bugs.  Right now the effect is to remove
Subdir.mk.tmp and main.mk.tmp, if they exist.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Fix breakage if nproc is not installed

This needs to echo 1, not try to run it.  The effect is a message on
stderr and passing `make -j'.  That is actually mostly OK since we
have few tests right now...

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system, tests: Marginally speed some things up with -f

This avoids an additional redirection through the stub Makefiles.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

comprehensive-test: Test the stest/ and mtest/ directories

Running these separately after make clean will help spot missing
dependencies.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

pretest-to-tested: Script for helping run comprehensive-test

This runs it on every commit between tested and pretest, advancing
tested as it goes.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

comprehensive-test: new script

This does an in-tree build, and two out-of-tree builds, one absolute,
and one relative.  It runs `make check' (not fullcheck, which is very
slow indeed).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest/udp-preload: Be more relaxed about paths

We would give ENAMETOOLONG if the specified path plus a maximal IPv6
address string wouldn't fit.  But in practice we don't use IPv6
addresses which stringify to anything that long, and this imposes an
annoyingly short restriction on the length of the emulation socket
directory path (a restriction which is going to be not met in my own
working tree when I add more sophisticated tests).

Instead, pass the suffix string to sun_prep and do the length check on
the actual length.  This means that addresses with short strings can
work, even if in principle addresses with long strings might fail.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest/udp-preload.c: Fix error handling of sun_prep

If sun_prep returned an error (eg, path too long), this would
segfault.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Cope if HEAD refers to a packed ref

In this case, there may be no corresponding .git/refs/<whatever>.
Simply tolerate this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Suppress MAKEFILE_TEMPLATES when running `make clean'

See the README hunk for the reasoning.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: break out subdirmk_target in generated Makefile

We are going to want to reuse this.  Make it a simply-expanded make
variable rather than a perl variable, so that it only gets computed
once.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: stub Makefiles: filter out `all'

Otherwise `make all' generates this:
  Makefile:3: target 'all' given more than once in the same rule
which is harmless but mildly irritating.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit '0428556765a15326025e9fe4809936c6cafd3c85'

subdirmk: clean.sd.mk: Delete .*~ too (eg .gitignore~)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: clean.sd.mk: Delete .*~ too (eg .gitignore~)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: cdeps.sd.mk: Honour &CDEPS_OBJECTS

This makes this facility more flexible.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: cdeps.sd.mk: Honour &CDEPS_OBJECTS

This makes this facility more flexible.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit 'a1e52614e600bed8e51b461e6569f2cd263e82bb'

subdirmk: Fix bug in regen.mk.in

Apparently, if you just write `;', make knows that the command can't
have done anything.  It therefore concludes tha the target file cannot
have been updated and doesn't bother rereading it.

The result is that you end up using last run's makefiles.

Writing @: fixes this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Fix bug in regen.mk.in

Apparently, if you just write `;', make knows that the command can't
have done anything.  It therefore concludes tha the target file cannot
have been updated and doesn't bother rereading it.

The result is that you end up using last run's makefiles.

Writing @: fixes this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit '6fce06f32c1dd21b5c2d7a142d54b7153d650574'

subdirmk: Change stub force target to `run-main.mk'

This produces significantly better error messages:
  [some error from actual makefiles]
  make[1]: Leaving directory '/home/ian/wherever'
  Makefile:6: recipe for target 'run-main.mk' failed

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Change stub force target to `run-main.mk'

This produces significantly better error messages:
  [some error from actual makefiles]
  make[1]: Leaving directory '/home/ian/wherever'
  Makefile:6: recipe for target 'run-main.mk' failed

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Fix bug when target in subdir exists

If you try to make a file that exists, the % rule doesn't seem to
match.  But providing the dependency for all of $(MAKCMDGOALS)
works.

We have to mention `all' explicitly because it might not appear in
$(MAKECMDGOALS).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Fix bug when target in subdir exists

If you try to make a file that exists, the % rule doesn't seem to
match.  But providing the dependency for all of $(MAKCMDGOALS)
works.

We have to mention `all' explicitly because it might not appear in
$(MAKECMDGOALS).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: start 0.5.2~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

stest: Suppress a GCC warning (for stretch)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

finalise 0.5.1

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

make-secnet-sites: Add test for -P option

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

make-secnet-sites: Actually fix -P option (!)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

mtest: Provide for sedderyb

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Release checklist: mention Subdir.sd.mk

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: Prepare 0.5.1 (more)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: Prepare 0.5.1

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit 'd817b49007e60b0e7392f23caf08dc0328eb32de' into HEAD

build system: Add warning suppressions for GCC9

We suppress 4 instances like:

  udp.c:113:45: error: increment of a boolean expression [-Werror=bool-operation]
    113 |     if (us->experienced[!!dest][af][success]++)
|                                             ^~

This is a very convenient idiom (using the saturating property of
boolean addition) and there is no sensible replacement.

And we suppress 6 instances like:

  tun.c:322:6: error: 'strncpy' specified bound 16 equals destination size [-Werror=stringop-truncation]
    322 |      strncpy(ifr.ifr_name,st->interface_name,IFNAMSIZ);
|      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

All write to ifr_name.  strncpy is precisely right for this, since the
API supports non-null-terminated names of length IFNAMSIZ.  That is
why I used strncpy.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

polypath.c: Fix a compiler warning

  polypath.c:520:20: error: '*' in boolean context, suggest '&&' instead [-Werror=int-in-bool-context]
    520 |     *allreasonable *= reasonable;
        |                    ^~

This is not really a very helpful suggstion because there is no &&=.
Happily the compiler doesn't (currently) complain about &= which is
the same when both arguments are bools.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

make-secnet-sites: sort the properties in level.output_props

This makes the test suite less sensitive to python version.
Now it works with 2.7.13, 3.5.3,, 3.7.5.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

make-secnet-sites: sort the properties

This makes the test suite less sensitive to python version.
Now it works with 2.7.13, 3.5.3.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

mtest: diff: Write output files and seddery rune

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

tests: Introduce `diff' proc

No functional change.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: tests: Add missing standard clean

We want (i) recursion from the parent (ii) automatic cleaning of
&CLEAN (which is set by cdeps.sd.mk, amongst other things).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Drop `check' from `all'

Otherwise there's no way to make everything but not run tets.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Do always regenerate version.o

We make two changes which affect only the litle recursive make of
version.o.  We only build version.o in the little sub-make when
relinking and then we want to always do it.  So the setting of OBJECTS
is pointless, and instead we make version.c depend on FORCE.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Fix nits following switch to subdirmk

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Docs: Fix some leftover ; and ,

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Docs, go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Docs, go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Implement new syntax

But I plan to go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Implement new syntax

But I plan to go back to ^ and ~

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit 'c0b216c79763f720df20e002fc14c9348a0e05c2'

docs: Discuss variable settings and inclusion order

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

subdirmk: Actually clear right variable

"subdirmk: main.mk: clear SUBDIRMK_MAKEFILES MAKEFILE_TEMPLATES"
cleared MAKEFILES by mistake (this was a rebase transposition error -
this commit originally came before "subdirmk: Rename MAKEFILES
variable".

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Abolish all absolute pathname syntaxes

These are rare and make things much more confusing.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Improve docs.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Fill in an empty cell with a copy instead

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Transpose columns in path syntax reference

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Change . to require rather than imply @

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Further corresponding new syntax

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: wip new syntax

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Swap rows in syntax table

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: New plan for directory references, further

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: New plan for directory references, new

Not yet implemented.  About to simplify this...

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: New plan for directory references

Not yet implemented.  About to simplify this...

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Summarise directory reference syntaxes, permute

Just permute the columns which seems to be clearer.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

docs: Summarise directory reference syntaxes

This reveals it to be a bit of a mess.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

legal: Include DEVELOPER-CERTIFICATE

A symlink to subdirmk's copy.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Clarify incremental paragraph

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Clarify race lossage paragraph

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Some hints in docs

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Honour CHECK_SILENT='' to print all the runes

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

build system: Fix out of tree builds again

"build system: convert stest and mtest to Subdir.mk" broke this by
referring to the wrong paths.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

CREDITS: Credit Mark properly (!)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Merge commit 'ff11733b6b7d7920033f81c4669ebba9131b984f' into wip.subdirmk

build system: ipaddrset.confirm

Make it like the others.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>