This code tries to prevent a large number of possible errors by
enforcing different restrictions on the messages that different
modules publish and subscribe to.

Some of these rules are probably too strict, and some too lax: we
should feel free to change them as needed as we move forward and
learn more.

This module implements a way to send messages from one module to
another, with associated data types.  It does not yet do anything to
ensure that messages are correct, that types match, or that other
forms of consistency are preserved.

We now accumulate warning flags in a separate variable,
"TOR_WARNING_FLAGS", and write it to a "warning_flags" file.  Then
we test whether the compiler will accept "@warning_flags": if so, we
put "@warning_flags" in the CFLAGS; if not, we copy the contents of
"$TOR_WARNING_FLAGS" into the CFLAGS.

Closes ticket 28924.

Alexander Hansen Færøy authored 6 years ago and Nick Mathewson committed 6 years ago

This patch removes the old process_handle_t code. Everything should by
now be using the process_t interface.

See: https://bugs.torproject.org/28179

Alexander Hansen Færøy authored 6 years ago and Nick Mathewson committed 6 years ago

This patch adds test cases for process_t which uses Tor's main loop.
This allows us to test that the callbacks are actually invoked by the
main loop when we expect them.

See: https://bugs.torproject.org/28179

This is a very gentle commit that just lays the groundwork in the
build system: it puts the include files to build libtor-app.a into
src/core, and to build the tor executable into src/app.  The
executable is now "src/app/tor".

(Term is short for terminal)

Note that procmon does *not* go here, since procmon needs to
integrate with the event loop.

Note that the workqueue code does *not* go here: it is logically at
a higher level, since it needs to use libevent and the networking
stack.

libtor-encoding is about various ways to transform data to and from
character sequences.

There are some additional changes to come: those points are marked
by XXXX.

The smartlist_core library now contains only the parts of smartlists
that are needed for the logging library.  This resolves the
circularity between "container" and "log".

The "containers" library still uses the logging code, and has the
higher-level smartlist functions.

The locking code gets its own module, since it's more fundamental
than the higher-level locking code.

Extracting the logging code was the whole point here. :)

The other time stuff is higher-level

I am calling the crypto library "crypt_ops", since I want
higher-level crypto things to be separated from lower-level ones.
This library will hold only the low-level ones, once we have it
refactored.

Apparently it has no testing variant.

We need this trick because some of our Rust tests depend on our C
code, which in turn depend on other native libraries, which thereby
pulls a whole mess of our build system into "cargo test".

To solve this, we add a build script (build.rs) to set most of the
options that we want based on the contents of config.rust.  Some
options can't be set, and need to go to the linker directly: we use
a linker replacement (link_rust.sh) for these.  Both config.rust and
link_rust.sh are generated by autoconf for us.

This patch on its own should enough to make the crypto test build,
but not necessarily enough to make it pass.