cpp-common/bt2c: add `DataLen` class (general-purpose data length)

This patch adds the `bt2c::DataLen` class so that you can make sure your
function accepts an expected data length instead of specifying that you
need bits or bytes:

    Something buildSomething(bt2c::DataLen len, ...)
    {
        using namespace bt2c::literals::datalen;

        if (...) {
            len += 16_MiBytes;
        } else {
            len *= 2;
        }

        // ...
    }

    // ...

    const auto something = buildSomething(DataLen::fromBytes(sz), ...);

As you can see above, you can use the `bt2c::literals::datalen`
namespace to access handy data length user literals.

The `bt2c::DataLen` constructor is private: you need to use the
bt2c::DataLen::fromBits() and bt2c::DataLen::fromBytes()
static methods to build an instance.

Because a `bt2c::DataLen` instance is just an `unsigned long long`
value in memory, the intention of this API is to copy data length
instances as much as possible. Therefore, I suppose it's cheaper to
accept a `const bt2c::DataLen` parameter than
`const bt2c::DataLen&` which would add an indirection.

The constructor and many methods are `constexpr` to make this useable in
constant context.

Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Change-Id: I1375fb5697c0d952f3669dc587603839d511dd18
Reviewed-on: https://review.lttng.org/c/babeltrace/+/7726
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12250

ctf: add ClkClsConfig and use it

Different parts of the code in the ctf plugin (in fs-src, lttng-live and
common) use these three values:

 - clock offset (seconds)
 - clock offset (nanoseconds)
 - force origin to be Unix epoch

Gather these three values in a new structure, `ClkClsConfig`, so that
these they can easily be passed around.  This replaces the existing
ctf_fs_metadata_config structure, which contained the exact same thing,
but was specific to the fs component class.

Change-Id: If9b4d88c9fd5eb3352c5de4007dc4343fac485c1
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8228
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12276

src.ctf.fs: remove uneeded check in create_streams_for_trace

At this point, we certainly have a stream class.

The ir_sc field is set in ctf_stream_class_to_ir.
ctf_stream_class_to_ir is called by ctf_trace_class_translate, which is
called by ctf_visitor_generate_ir_visit_node if ctx->trace_class is set.
That trace_class field is set if a self_comp was passed to
ctf_visitor_generate_ir_create, in the decoder_config object.  Going up
the stack, we get to ctf_fs_metadata_set_trace_class, called by
ctf_fs_trace_create, called by ctf_fs_component_create_ctf_fs_trace.

create_streams_for_trace is called by ctf_fs_trace_create, which always
passes a self_comp to ctf_fs_component_create_ctf_fs_trace, ensuring we
always end up with a stream class.

ctf_fs_component_create_ctf_fs_trace is also used in the context of
queries, where no self_comp if passed, and therefore no stream class is
created..  But in that context, create_streams_for_trace is not called.

Change-Id: I5e8ef72a1013fc156a7bc6ac0d4cbb231a5ab704
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8237
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12274
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: query.cpp: change assertions into exceptions

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Change-Id: I21d5f8950ca8efa78713404a29c36442cba6502a
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8181
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12273
Tested-by: jenkins <jenkins@lttng.org>
CI-Build: Simon Marchi <simon.marchi@efficios.com>

cpp-common: add `dataFromFile()`

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Change-Id: Ief289d7d12a1cc964af16dc7dd74a4422c81920a
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8159
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12272
CI-Build: Simon Marchi <simon.marchi@efficios.com>
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: use C++ value wrappers in queries

 - Pass the query parameters as a bt2::ConstMapValue
 - Return the query result as a bt2::Value::Shared
 - Communicate errors via exceptions
 - Use std::string instead of GString to build the metadata text

Change-Id: If75da1efc1accb5c42dd98d830ab9abf2ffff932
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8165
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12271
Tested-by: jenkins <jenkins@lttng.org>

ctf: use unique_ptr to manage ctf_metadata_decoder lifetime

Introduce the ctf_metadata_decoder_up type, a unique_ptr with a deleter
that calls ctf_metadata_decoder_destroy.  Change
ctf_metadata_decoder_create to return a ctf_metadata_decoder_up, and
adjust callers / callees accordingly.

Note that this is temporary, ctf_metadata_decoder is going to be deleted
in a future refactor.  But in the mean time, it helps make the callers
use RAII and become more exception-safe.

Change-Id: Ia0e24b425c47b90dc71b10d1c7fd0a3000c89180
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8167
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12270
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: manage lifetime of open file with FileUP

Change-Id: Ic1c5d8be3761e449a154e1fd8cc1f0ec7c957197
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8173
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12269
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: use GCharUP to hold metadata path

Change-Id: I2cf43f7da567da3af2786a87026e25272925ed6f
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8171
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12268
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: make ctf_fs_make_port_name return a GCharUP

Make ctf_fs_make_port_name return a GCharUP, which then helps use
automatic memory management in the callers.

Change-Id: I9191ddbafa86a4788fc5e16fc561dfda7696cfa6
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8166
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12267
Tested-by: jenkins <jenkins@lttng.org>

src.ctf.fs: use unique_ptr to manage ctf_fs_component lifetime

Define ctf_fs_component::UP to be a unique_ptr with a deleter that calls
ctf_fs_destroy.  Change ctf_fs_component_create to return a
ctf_fs_component::UP and adjust the appropriate callers / callees.

Move where the ctf_fs_component instance is assigned as the "data" of
the self component.  Make it so it's assigned in ctf_fs_init (the
component initialization entry point), as the very last step.  This way,
ctf_fs_create can return a unique_ptr, and there's no ambiguity on who
owns the ctf_fs_component instance at any given time, and what happens
in case of failure.

Change-Id: I3ab1d70b0e7aa7772b7fa09deec70db7ea5a022c
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8164
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12266
Tested-by: jenkins <jenkins@lttng.org>

ctf: use unique_ptr to manage ctf_visitor_generate_ir lifetime

Define ctf_visitor_generate_ir::UP as a unique_ptr type with a custom
deleter that calls ctf_visitor_generate_ir_destroy.  Make
ctf_visitor_generate_ir_create return a ctf_visitor_generate_ir::UP and
adjust relevant callers / callees.

Change-Id: I9db22bd66e69473e67c2571e0b2bb7682863cff7
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8163
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12265
Tested-by: jenkins <jenkins@lttng.org>

ctf: move ctf_visitor_generate_ir to header file

Move ctf_visitor_generate_ir to the `ast.h` header file, the header file
containing all the declarations for the CTF 1 metadata visitors.  This
will allow exposing it as a proper C++ class and remove the accessor
functions.

Change-Id: Ic13494d0040ee9e2214ecca45d8ea2e7b75ec4fa
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8162
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12264
Tested-by: jenkins <jenkins@lttng.org>

ctf: add try/catch at entry points

Add try / catch at the entry points of the CTF component classes to
convert thrown exceptions in the proper statuses.  The handled
exceptions are:

 - std::bad_alloc: converted to MEMORY_ERROR
 - bt2::Error: converted to ERROR

Any other exception (for example, coming from a third-party library we
would eventually use) should be caught and converted to a
bt2_common::Error.

Change-Id: I06c2ccb43947f5a9e61568967f2812ccb0821b36
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12263
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

sink.text.details: don't write empty strings

This patch makes a `sink.text.details` not write the value of a trace
environment entry or of a value object when it's an empty string.

This is mostly because I don't like to keep plain text files with
significant trailing whitespaces in this repository, and
`data/plugins/src.ctf.lttng-live/cli-multi-domains.expect` had some,
for example:

    trace_creation_datetime:️▒
    trace_name:▒

where `▒` is a space.

Of course we'll have the same issue if a user string contains only
whitespaces instead of being empty, but since it's unlikely that we'll
do this in our own tests, this patch is good enough for me.

Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Change-Id: I6be83e2721bc6ce98e1459a7df42e32ccf6b661d
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12249
CI-Build: Simon Marchi <simon.marchi@efficios.com>
Tested-by: jenkins <jenkins@lttng.org>

sink.ctf.fs: use bt2c::Logger

Use `bt2c::Logger` throughout the `sink.ctf.fs` code.

When printing C strings that may be NULL (e.g. result of
`bt_trace_get_name()`), we currently rely on glibc printf's behavior of
printing `(null)` when passed a NULL pointer for `%s`.  {fmt} doesn't
accept a nullptr `const char *`.  Add the `bt2c::maybeNull()` function
to emulate the existing behavior.

Change-Id: I6ab259c8db526a3b4c0de9295a4360180c56b398
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12241
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: make `FieldPathScope` a wise enum

Change-Id: Ia07b63ff3afe01401ba633d0787ed726143abb05
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12247
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: make `ValueType` a wise enum

Change-Id: I0a0b060a094022844b159cedba87de566d4cfbac
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12246
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

sink.ctf.fs: initialize structure fields

In preparation to using some C++ objects as fields.  Allocate with new
and deallocate with delete where relevant.

Change-Id: I924a6088b17e1d3ae4b8a239cf059fc3a7ecc8c6
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12239
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

ctf: rename `ctx` structures

Rename the various structures named `ctx` to something else, to avoid
ODR problems when C++-ifying them.

Change-Id: I1b30c38ba6b88687b5baa5b965b805b0a9097ffe
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12236
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: make `UniqueConstError::_mError` a pointer

Make this field a pointer instead of a reference, in order to make
`UniqueConstError` assignable.  I don't really have a need for this, but
it seems more correct.

Change-Id: I29ae7a7380f75fa6b680116168a309e70c381f11
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12235
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

ctf: use `bt2c::Logger` throughout `src.ctf.fs`, `src.ctf.lttng-live`

A lot of functions and structures have parameters or fields of type
`bt_self_component`, `bt_self_component_class` and `bt_logging_level`
for the sole purpose of logging.  This becomes quite verbose, especially
when a function can be used in both self component and self component
class context, then it will have both of these parameters.

The `bt2c::Logger` class exists to encapsulate this context in a single
value that can be passed around.  Use it as much as practically possible
(as well as the corresponding logging macros) throughout the sources of
the ctf plugin, as well as in the common directory that they depend on.

Update all format strings to use {fmt}'s format string syntax.

Change all "enum to string" functions to be `format_as()` functions
understood by {fmt}.  Move these functions near the type they format,
for convenience.

Remove all the macros that appear at the beginning of source files that
influence the logging (such as `BT_COMP_LOG_SELF_COMP` and `BT_LOG_TAG`).

The log tags were previously defined per-source file, whereas they are
now defined per `Logger` object.  I tried to re-use the existing log
tags when it made sense, but in some places the log tags are now a bit
more granular.

Remove the `ctf_plugin_metadata_log_level` variable, controlled by the
`BABELTRACE_PLUGIN_CTF_METADATA_LOG_LEVEL` environment variable.  This
variable controlled the logging level in the lexer and parser (generated
by flex and bison).   It existed because the generated lexer and parser
don't provide an easy way to pass user data around.

Work around this by define the following variable:

   thread_local const ctf_scanner *currentCtfScanner;

and making the logging in the lexer and parser code use
`currentCtfScanner->logger`.  Set `currentCtfScanner` when entering the
parser in `ctf_scanner_append_ast()` and install an RAII cleanup to
reset it.

Change-Id: I0f6844d2cb83d2bf52b81ae9dee4179c7b1f2e12
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12199
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2c: make `Logger::Level` a wise enum

Change-Id: I5298a6af1a6e5a8127d20f546746d5050d767f06
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12245
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: make `MessageType` a wise enum

Change-Id: I46182e774ae480b20461ce44add48cea47bb0f3a
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12244
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2c: add format_as() template functions for wise enums

Add some format_as() functions in the bt2 and bt2c namespaces that will
apply to all wise enums defined in that namespace.

Change-Id: I4952f062fc36e94925d2f0b093ac09d0fae6fdc0
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12248
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/vendor: add copy of wise_enum

Add a copy of wise_enum, a smart enum library that allows doing cool
things like getting the number of enumerators of an enum, iterating on
enumerators, and converting enumerators to string while effectively
keeping vanilla `enum`/`enum class` behind the scenes (no enum-like
class).

The version is 3.1.0, which corresponds to the git tag `3.1.0`, which
corresponds to this git commit:

  https://github.com/quicknir/wise_enum/commit/34ac79f7ea2658a148359ce82508cc9301e31dd3

Change-Id: I9b92d5cd012dee427cdd725c17eeae9f2a529cb0
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12242
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: move some enum classes out of enclosing classes

Settle on the standard of having these enumerations out of the classes
that use them.

Change-Id: I241bdca560bf7f9e892d8884c86662166ad2d2c4
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12240
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: use relative include for uuid.hpp in fmt.hpp

Change-Id: Ic5b5dee8cd967ab8bc5ab476747e4b09e7aa1127
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12237
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: add `BT_CPPLOGE_ERRNO_APPEND_CAUSE_SPEC`

Add the `BT_CPPLOGE_ERRNO_APPEND_CAUSE_SPEC` macro, allowing to log an
error and append an error cause based on `errno`, without throwing.

Change-Id: I5837ef4b78a6d89b5574072f23d4aba6665e6bb4
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12232
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: add `BT_CPPLOGE_APPEND_CAUSE_SPEC`

Add the `BT_CPPLOGE_APPEND_CAUSE_SPEC` macro, allowing to log an error,
append a cause, but not throw.

Change-Id: I6a09dd8775a85202c2049e902ab9edfcbd32080b
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12231
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2c: Logger: remove `noThrow` suffixes

I think that the `noThrow` suffixes are unnecessary.  Since the methods
that throw (or rethrow) have the `AndThrow` (or `AndRethrow`) suffixes,
it is implied that those that don't, don't.

Simplify some macro comments to omit "without appending nor throwing",
for the same reason (if the macro name doesn't contain "APPEND_CAUSE",
it doesn't append a cause, if it doesn't contain "THROW", it doesn't
throw).

Change-Id: Ia71cf733e857e89578648eda12bf64b2bfc65027
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12229
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: fix some function names in comments

Change-Id: I1d66d431e36794daf3b35ab5b3600bcea3bc42d1
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12225
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

Change naming convention for enum class enumerators

Change the naming convention for enum class enumerators to not use
`ALL_CAPS`, as suggested by the C++ Core Guidelines [1].  Instead, use
camel case starting with a capital letter.

Update `CONTRIBUTING.adoc` to reflect this change.

gcc emits warnings like this:

    /home/smarchi/src/babeltrace/src/cpp-common/bt2/trace-ir.hpp:880:9: error: declaration of 'bt2::CommonEventClass<LibObjT>::LogLevel::Error' shadows a global declaration [-Werror=shadow]
      880 |         Error = BT_EVENT_CLASS_LOG_LEVEL_ERROR,
          |         ^~~~~
    In file included from /home/smarchi/src/babeltrace/src/cpp-common/bt2/clock-class.hpp:20,
                     from /home/smarchi/src/babeltrace/src/clock-correlation-validator/clock-correlation-validator.cpp:7:
    /home/smarchi/src/babeltrace/src/cpp-common/bt2/exc.hpp:14:7: note: shadowed declaration is here
       14 | using Error = bt2c::Error;
          |       ^~~~~

Whether or not this is a compiler bug is not clear [2], but in our case
we know it's fine.  Add and use `BT_DIAG_IGNORE_SHADOW` in some places
to silence these warnings.

[1] https://cpp-core-guidelines-docs.vercel.app/enum#Renum-caps
[2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55776

Change-Id: I59bb7f3c8572106c325596bc80a4d70714d86dee
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12212
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

Fix: cpp-common/bt2c: copy `_mSelfCompCls` when constructing `Logger`

Change-Id: Ic7e8ddc09a4f943f845821e16323cf2d5c1a667c
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12211
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

build: make out-of-tree builds with Python bindings work on msys2

When doing an out of tree build on msys2 with
`--enable-python-bindings`, where the path to configure is absolute, I
get:

    Making all in bindings/python/bt2
    make[3]: Entering directory '/home/smarchi/build/babeltrace-mingw-w64-ucrt-x86_64/src/bindings/python/bt2'
    touch copy-static-deps.stamp
      SWIG      bt2/native_bt.c
    CC="ccache gcc" CFLAGS="-IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0/include -fvisibility=hidden -pthread -IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0/include   -Wall -Wextra -Wmissing-prototypes -Wmissing-declarations -Wnull-derefere
nce -Wundef -Wredundant-decls -Wshadow -Wjump-misses-init -Wsuggest-attribute=format -Wnested-externs -Wwrite-strings -Wformat=2 -Wstrict-aliasing -Wmissing-noreturn -Winit-self -Wduplicated-cond -Wduplicated-branches -Wlogical-op -Wsuggest-override -Wno-sign-compare -Wno-cast-functi
on-type -Wno-maybe-uninitialized -Wold-style-definition -Wstrict-prototypes -g -O2 -Wno-shadow -Wno-null-dereference -Wno-deprecated-declarations -Wno-redundant-decls -Wno-missing-field-initializers -Wno-unused-parameter -Wno-undef" CPPFLAGS="-I/home/smarchi/src/babeltrace/include -I
../../../../src -I/home/smarchi/src/babeltrace/src -include common/config.h  -I/home/smarchi/src/babeltrace/src/bindings/python/bt2/bt2" LDFLAGS="-L../../../../src/lib/.libs  -lgmodule-2.0 -lglib-2.0 -lintl -LC:/msys64/ucrt64/lib -lpython3.11  -lversion -lshlwapi -lpathcch -lbcrypt -
lm   -lgmodule-2.0 -lglib-2.0 -lintl" /ucrt64/bin/python ./setup.py build --force
    [...]
    ccache gcc -IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0/include -fvisibility=hidden -pthread -IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0/include -Wall -Wextra -Wmissing-prototypes -Wmissing-declarations -Wnull-dereference -Wundef -Wr
edundant-decls -Wshadow -Wjump-misses-init -Wsuggest-attribute=format -Wnested-externs -Wwrite-strings -Wformat=2 -Wstrict-aliasing -Wmissing-noreturn -Winit-self -Wduplicated-cond -Wduplicated-branches -Wlogical-op -Wsuggest-override -Wno-sign-compare -Wno-cast-function-type -Wno-ma
ybe-uninitialized -Wold-style-definition -Wstrict-prototypes -g -O2 -Wno-shadow -Wno-null-dereference -Wno-deprecated-declarations -Wno-redundant-decls -Wno-missing-field-initializers -Wno-unused-parameter -Wno-undef -IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0
/include -fvisibility=hidden -pthread -IC:/msys64/ucrt64/include/glib-2.0 -IC:/msys64/ucrt64/lib/glib-2.0/include -Wall -Wextra -Wmissing-prototypes -Wmissing-declarations -Wnull-dereference -Wundef -Wredundant-decls -Wshadow -Wjump-misses-init -Wsuggest-attribute=format -Wnested-ext
erns -Wwrite-strings -Wformat=2 -Wstrict-aliasing -Wmissing-noreturn -Winit-self -Wduplicated-cond -Wduplicated-branches -Wlogical-op -Wsuggest-override -Wno-sign-compare -Wno-cast-function-type -Wno-maybe-uninitialized -Wold-style-definition -Wstrict-prototypes -g -O2 -Wno-shadow -W
no-null-dereference -Wno-deprecated-declarations -Wno-redundant-decls -Wno-missing-field-initializers -Wno-unused-parameter -Wno-undef -IC:/msys64/home/smarchi/src/babeltrace/include -I../../../../src -I/home/smarchi/src/babeltrace/src -include common/config.h -I/home/smarchi/src/bab
eltrace/src/bindings/python/bt2/bt2 -IC:/msys64/ucrt64/include/python3.11 -c /home/smarchi/src/babeltrace/src/bindings/python/bt2/bt2/logging.c -o build/temp.mingw_x86_64_ucrt-3.11/home/smarchi/src/babeltrace/src/bindings/python/bt2/bt2/logging.o
    cc1.exe: warning: command-line option '-Wsuggest-override' is valid for C++/ObjC++ but not for C
    cc1.exe: fatal error: /home/smarchi/src/babeltrace/src/bindings/python/bt2/bt2/logging.c: No such file or directory
    compilation terminated.
    error: command 'C:\\msys64\\ucrt64\\bin/ccache.exe' failed with exit code 1
    make[3]: *** [Makefile:735: build-python-bindings.stamp] Error 1
    make[3]: Leaving directory '/home/smarchi/build/babeltrace-mingw-w64-ucrt-x86_64/src/bindings/python/bt2'

For 'in-tree' builds all the paths are relative and thus compatible with
native windows binaries while in an 'out-of-tree' build some of the
paths are absolute and thus incompatible. The MSYS2 shell does a lot of
'magic' conversion [1] on paths passed as arguments and through environment
variables but the heuristics are limited and it doesn't apply when
binaries are called directly without using the shell.

The first issue is the include paths in the CFLAGS and CPPFLAGS
environment variables which don't get converted because they start with
'-I'.

The second issue is that the path to 'logging.c' in the generated
'setup.py is absolute:

    sources=["bt2/native_bt.c", "/home/smarchi/src/babeltrace/src/bindings/python/bt2/bt2/logging.c"],

When distutils invokes gcc to compile this file, it spawns it directly,
not through a shell, so no automatic Unix -> Windows path conversion
takes place.

Fix both of those issues by modifying 'setup.py' to detect the MSYS2
environment and convert absolute path to Windows style using the
'cygpath' command.

[1] https://www.msys2.org/docs/filesystem-paths/

Change-Id: I9003878a6582a0dc3f4ee18b07885f332348b8bf
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12397
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Simon Marchi <simon.marchi@efficios.com>

src/Makefile.am: build {fmt} without `-Wmissing-noreturn`

Until <https://github.com/fmtlib/fmt/pull/3917> is merged and becomes
part of a release, let's just ignore missing `[[noreturn]]` warnings
when building {fmt}.

Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Change-Id: Ia63f6248a59d8b42b35523e0a1972daa94f957bf
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12223
Reviewed-by: Simon Marchi <simon.marchi@efficios.com>
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Michael Jeanson <mjeanson@efficios.com>

Fix: cpp-common/bt2c/logging.hpp: remove one definition of `BT_CPPLOGE_APPEND_CAUSE_AND_THROW_SPEC`

`BT_CPPLOGE_APPEND_CAUSE_AND_THROW_SPEC` has two identical definitions,
remove one.

Change-Id: Ia5b7caae53b65169081e9d8214896887fa062ad4
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12198
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

Fix: cpp-common/bt2c/logging.hpp: don't set `_mSelfComp` when creating `Logger` for message iterator

Only one context in `Logger` must be set.  Similarly, when creating a
`Logger` for a component, we don't set `_mSelfCompCls`.

Change-Id: Ie826986d301cb996b594127c72094b90974b8e87
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12197
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

compat: rename socket.h to socket.hpp

`socket.h` is only included in C++ code, so make it a C++ file.  This
makes it possible for an upcoming patch to make it use `bt2c::Logger`.

Change-Id: I42c9fd08a343357c116cf85675f36997cc1f3f2e
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12196
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

ctf: allocate some structures with new

Allocate some key structures in the ctf plugin with new (and free with
delete), so we can start using non-POD fields in them.  Initialize all
fields to 0 / nullptr to replicate what g_new0 did.

Add ctf_fs_ds_index_entry_destroy, to abstract how this object is
destroyed (it is currently being g_free'd at multiple places) and to use
it as a g_ptr_array_new_with_free_func callback.

Change-Id: Ic0f893655db22c964641639aa0625d820af71587
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8021
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12195
Tested-by: jenkins <jenkins@lttng.org>

ctf: remove unused `struct ctf_metadata_decoder`

I noticed this struct was unused.  There is another `struct
ctf_metadata_decoder` in src/plugins/ctf/decoder.cpp, that is actually
used.

Change-Id: Ief9a4e37e7414b8fcd729581bb040c1100722327
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/8020
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12191
Tested-by: jenkins <jenkins@lttng.org>

src/plugins/ctf/common: restructure subtree

This patch restructures the `src/plugins/ctf/common` subtree as such:

`src`:
    Everything common to source component classes.

`src/metadata/tsdl`:
    TSDL-specific code.

    As of this patch, everything which was part of
    `src/plugins/ctf/common/metadata` is moved here, but some code will
    be shared with CTF 2 sources eventually and move back to
    `src/plugins/ctf/common/src/metadata`.

`print.hpp` isn't needed anymore.

This new structure will make it easier to add common `ctf` plugin code
for both source and sink component classes, as well as to add code
specific to CTF 2 and common to both CTF 1 and CTF 2.

Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Change-Id: I4246d6ba37ff26efa5d4b7b4be2ec8600d3207e4
Reviewed-on: https://review.lttng.org/c/babeltrace/+/7839
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12190
CI-Build: Simon Marchi <simon.marchi@efficios.com>
Tested-by: jenkins <jenkins@lttng.org>

sink.text.pretty: abort on unknown field class type

I don't think there a legimitate case for the print_field function to
encounter an invalid field class type.  Change it to abort instead of
returning an error.

Change-Id: I333d2221b828841aff38a7d8b0a17032af3aafb1
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/7580
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/7573

sink.ctf.fs: remove unused `fs_sink_ctf_field_class::index_in_parent` field

This field is not used anywhere and looks like a leftover from a
previous iteration of the sink. Let's remove it.

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Change-Id: I64592c1561bd66e083e965903d8a9dc050a5756d
Reviewed-on: https://review.lttng.org/c/babeltrace/+/7387
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12189
Tested-by: jenkins <jenkins@lttng.org>
CI-Build: Simon Marchi <simon.marchi@efficios.com>

configure.ac: build with `-DFMT_EXCEPTIONS=0`

Coverity flags some functions marked `noexcept` that call `fmt::format`,
since `fmt::format` can throw `fmt::format_error` exceptions.  We're not
interested in handling format errors at runtime, we'd rather just get a
crash and fix it.  Build the entire project (including `libfmt.la`) with
`-DFMT_EXCEPTIONS=0` to make {fmt} crash instead of throw, should a
format error occur.

Tested by introducing a format error somewhere and running to it.  The
result is very pretty:

    $ tests/plugins/flt.utils.muxer/test-clock-compatibility
    1..150
    /home/smarchi/src/babeltrace/src/cpp-common/vendor/fmt/format-inl.h:40: assertion failed: invalid format stringterminate called without an active exception
    [1]    1189949 IOT instruction  tests/plugins/flt.utils.muxer/test-clock-compatibility

Change-Id: Ibcb0104a6bf644b025112d0b35e4fbe622c20be9
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12194
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Reviewed-by: Michael Jeanson <mjeanson@efficios.com>
Tested-by: jenkins <jenkins@lttng.org>

Fix: bt_value_map_borrow_entry_value(): use `@bt_pre_is_map_val`

Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Change-Id: I48f15e9938d78245fd5f86a35935bf56fd304261

.clang-tidy: enable cppcoreguidelines-avoid-const-or-ref-data-members

Enable:

  https://clang.llvm.org/extra/clang-tidy/checks/cppcoreguidelines/avoid-const-or-ref-data-members.html

Set the `AllowMissingMoveFunctionsWhenCopyIsDeleted`, so it's less
verbose when you just want to delete everything.

Set `AllowSoleDefaultDtor`, because we have legitimate cases (when
defining virtual base classes) where we only define a default
destructor.  Perhaps there's a pitfall I don't see when allowing this?

Change-Id: I4a8d20c0be628ac807db4c6b52548e3b7d7ee286
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12090
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

Add clang-tidy config file

Copied from:

  https://git.lttng.org/?p=lttng-tools.git;a=blob;f=.clang-tidy;h=75f559c5d961041df8e37280ef27cb0cac8cdd96;hb=20c4b46aee1a984bc558b483826c1078f24d35e9

And adapted for Babeltrace:

 - Define `BT_ASSERT` and `BT_ASSERT_DBG` as assertion macros.
 - Remove `HeaderFilterRegex`.  We might want to use this setting, but
   I'm not sure what its value would be for Babeltrace.
 - Fix a . -> , typo

Change-Id: I73f1776bc4b7cb1bdc0d5728f9a46b17157088c8
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12078
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

flt.utils.muxer: use clock correlation validator

Change the clock correlation validation code in `flt.utils.muxer` to use
the clock correlation validation library.

Change-Id: I5452f530de5fb035f66da5fc6f63bba48711e503
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11991
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

flt.utils.muxer: don't reset clock expectation when seeking beginning

When seeking the beginning, there is no reason to reset the clock
expectations.  The clocks seen after a seek beginning should be
compatible with those seen before.

Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Change-Id: I2ea587ea7b7f62ae509632836ae4364292903001
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12148
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: add `format_as()` functions for `Uuid` and `UuidView`

Change-Id: I6530144e78d6aa10dedb320b90a4c208aa572e9d
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12138
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

tests/plugins/flt.utils.muxer: add test for clock (in)compatibility cases

Add a test for the various clock class incompatibility cases checked by
the `MsgIter::_makeSureClkClsIsExpected` method in the `flt.utils.muxer`
component class.

The scenarios are equivalent to the triggers in
`tests/lib/conds/conds-trigger.cpp`, used to test the preconditions in
the library.  Since the scenarios are equivalent, it would probably be
possible to share some code with that file, but I haven't tried it.

In essence, for each scenario, the test creates a graph where a muxer
component receives two messages with incompatible clock
configurations.  The test expects that an exception is thrown, and
checks whether the root cause contains an expected string.

Change-Id: I8bd98fa52719ab2cedb7cbb09310a560b437eb28
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12005
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

plugins/common/muxing: compare "origin is Unix epoch" property of clock classes

If two clock classes only differ by their "origin is Unix epoch"
property, they will compare equal as far as `compare_clock_classes` is
concerned.  Add a comparison for that.

Change-Id: Iecac8fd4aa3064b221291264f4fc6758e32a0943
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12067
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

plugins/common/muxing: compare stream class clock classes

When comparing two streams and all else is equal about the stream and
their classes, compare the clock classes attached to the stream classes.

My use case for this is that I'm writing a test where two source
components send a stream beginning message to a muxer components.  The
only difference between the two is the clock classes associated to the
stream classes.  Some stream classes have an associated clock class,
some don't.  In other cases, the stream classes differ only by some
properties of their clock classes.  I want to be sure that the messages
will come out of the muxer's priority heap in a stable order.

Change-Id: I113704dac4ff952573c3dd96acd33b295425a863
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12054
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

plugins/common/muxing: split `compare_clock_snapshots_and_clock_classes()`

A subsequent commit will want to re-use the code to compare two clock
classes.  Split `compare_clock_snapshots_and_clock_classes()` in two, to
make the comparing clock classes part independent from the comparing
clock snapshots part.

Change-Id: I7888d7d3d1339ccb6c27054919349745cb7875eb
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12053
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

sink.utils.counter: remove unnecessary check in counter_consume()

Given the lifecycle of a component, I'm pretty sure this condition can
never be true.  For counter_consume() to be called,
counter_graph_is_configured() must execute successfully.  If
counter_graph_is_configured() executes successfully, `counter->msg_iter`
is necessarily set.

Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Change-Id: Ibd3cbcfdeefa8d79143e39c6fb846ca5eeebd249
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12010
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

sink.utils.dummy: remove unnecessary check in dummy_consume()

Given the lifecycle of a component, I'm pretty sure this condition can
never be true.  For dummy_consume() to be called,
dummy_graph_is_configured() must execute successfully.  If
dummy_graph_is_configured() executes successfully, `dummy->msg_iter` is
necessarily set.

Change-Id: I5c11c6b76c8cb267f0185c1a35bc2f5e70826509
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12009
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

sink.utils.counter: append error cause when call to bt_message_iterator_next()` fails

Same as the previous patch, but for the counter component.

Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Change-Id: I95f27091512aebc9d01c7b05d98955fdfdcab5da
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12008
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

sink.utils.dummy: append error cause when call to bt_message_iterator_next()` fails

I'm looking at some error stacks for a graph involving a dummy
component, and I noticed that it did not produce an error cause when
failing to consume.  I think it would be nice if it did, to avoid having
a gap in the error stack.

Change-Id: I1b7af3218f9937a21d3dee7406a2b03e122671c6
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12007
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: use universal reference for `Graph::addComponent()` initialization data

I noticed it was not possible to pass data to `Graph::addComponent`
using a temporary object:

    graph->addComponent(*srcCompCls, "source-1", TestSourceData {});

My understanding of it is that the compiler does not consider the
overload with the `InitDataT&` parameter for this, because the temporary
object can't bind to the lvalue reference.

Change these methods to take forwarding references instead.

Change-Id: Icc80239812c325b777ed83e8634302a0c9d72f14
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12049
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

Fix: flt.utils.muxer: reject two different clock classes with unknown origin and no UUID

When getting a message with a clock class whose origin is unknown and
that has no UUID, the expectation of a `flt.utils.muxer` component
should be that all subsequent messages with an associated clock class
will have the exact same clock class instance.  Currently, it accepts
any other clock class instance that has an unknown origin and no UUID.

This is a change that is equivalent to what was done in commit
29e191fceb61 ("Fix: lib: strengthen clock expectation check for no Unix
epoch / no UUID case") in the library iterator code.

When first seeing a clock class with unknown origin and no UUID, make
the muxer message iterator save the clock class instance, holding a
strong reference, and ensure that all clock classes seen after are the
same instance.

Change-Id: I676471cc0c3b88a83cd0b452dee8d6b1ed549e4d
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12044
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: add `CStringView::startsWith`

Change-Id: I4a39561085850c06adc1f6e36466fba9a2e6dfa0
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12011
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: add C++ bindings around `bt_error *`

Add C++ bindings to access the error on the current thread.

 - Take the current thread error with takeCurrentThreadError(),
   returning a `UniqueConstError`.
 - Move the current thread error back with moveErrorToCurrentThread().
 - Access the error cause by iterating or subscripting
   `UniqueConstError`, which yields objects of type `ConstErrorCause` (a
   borrowed object type).
 - Access the common error cause information using methods of
   `ConstErrorCause`, including the error cause actor type.
 - Obtain a more specific error cause object with
   `ConstErrorCause::as*()`.

Change-Id: Ie18e22ef8901bc9c8d7c85432d515ca763b484c2
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12006
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

lib/graph/iterator.c: fix indentation (spaces -> tabs)

Change-Id: Ic6b408ef73e54e59161dca18fe19a45285d39c22
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12123
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

lib/graph/iterator: use clock correlation validator util

Make `graph/iterator.{c,h}` use the clock correlation validator
convenience library.

 - Replace the fields of `struct bt_message_iterator` that track the
   clock expectation state with a single `struct
   bt_clock_correlation_validator` field, only present if babeltrace is
   built in dev mode.

 - Initialize `bt_message_iterator::correlation_validator` in
   create_self_component_input_port_message_iterator.  Finalize it in
   `bt_message_iterator_destroy`.

 - Replace the bulk of
   `assert_post_dev_clock_classes_are_compatible_one` with one call to
   `bt_clock_correlation_validator_validate_message`.  If the latter
   returns an error, trigger an assertion failure based on the return
   error type and associated information.

The modifications in the Makefiles (adding `dummy.cpp`) are to force
linking with the C++ compiler driver.  This is necessary, since the
`libbabeltrace2.so` library now requires some symbols from the C++
standard library (e.g. `libstdc++.so`).  The `dummy.cpp` does not
actually need to exist.  This trick is described in the automake doc
[1].

[1] https://www.gnu.org/software/automake/manual/automake.html#Libtool-Convenience-Libraries

Change-Id: I46a3a553d5e18eb6860d65d928ef33d941e9e951
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11990
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

lib/graph/iterator: don't reset clock expectation when seeking beginning

When asking an iterator to seeking the beginning, there is no reason to
reset the clock expectations.  The clocks seen after a seek beginning
should be compatible with those seen before.

Change-Id: I91d45b22c477048d92bc567a3ee26ec9641c1055
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12147
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

clock-correlation-validator: add clock correlation validation util

`lib/graph/iterator.c` and `plugins/utils/muxer/msg-iter.cpp` have
equivalent code to verify that the various clock classes they see are
correlatable.  Add a util to factor out this logic.

Since the lib is written in C and the muxer component in C++, we need
both interfaces.  I chose to write the core of the library with a proper
C++ interface (since that's the future!) and make a small C wrapper
around it.  If/when we migrate some parts of the lib to C++, we will be
able to drop this C wrapper.

Change-Id: Iab8968ef33c9f7d4e2e8e06e0942de749a090841
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11753
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: add `wrap(bt_graph *)`

Change-Id: I85e6e38576d71f64a7cd8cb0139104bce98aac36
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12046
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: don't use wrap in bindings code

Using `wrap()` in bindings code causes cyclic include problems, stop
using it.

For instance, we want to add `wrap(bt_graph *)`, so `wrap.hpp` needs to
include `graph.hpp`.  But `graph.hpp` currently uses and includes
`wrap.hpp`.

Change-Id: I5b7005617ed38eb791edf5802a05a8a8dd3115fd
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12045
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: make `getSupportedMipVersions` methods take a `ConstMapValue`

The "get_supported_mip_versions" component class methods always receive
a map value.  The equivalent methods in the C++ bindings should
therefore take a `ConstMapValue.

Change-Id: Ia3ba17048337d3abb72cefa2ae88c7b3187833f2
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12048
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: remove unnecessary mentions of `bt2::`

Don't use `bt2::` when already inside the `bt2` namespace.

Change-Id: I63444e141ae13b6e508d4508488fb562ae9403d7
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12047
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: make `{OptionalBorrowedObject,SharedObject}::operator bool` explicit

This code doesn't really make sense (what I really wanted to do was to
compare the underlying lib pointers), but I made this mistake on more
than one occasion:

    if (someOptionalBorrowedObject == someSharedObject)

What happens is: since `operator bool` for both types is implicit, they
get implicitly converted to bool.  As a result, the condition checks if
both objects contain an object or don't contain an object.  I consider
this a pitfall, as it compiles but it is likely never what we'll want to
do.

Make `operator bool` for both types explicit to avoid that.  This is
what is done in the STL for `shared_ptr` [1] and `optional` [2].

Note that even if `operator bool` is explicit, it remains implicitly
convertible to bool in certain contexts [3], such as an if condition.

[1] https://en.cppreference.com/w/cpp/memory/shared_ptr/operator_bool
[2] https://en.cppreference.com/w/cpp/utility/optional/operator_bool
[3] https://stackoverflow.com/questions/39995573/when-can-i-use-explicit-operator-bool-without-a-cast

Change-Id: I2b2377323ca620162442cb7f0305763c034d13d8
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12043
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

flt.utils.muxer: use `streamClassDefaultClockClass()` methods

Use them where possible, as a shortcut to
`.stream().cls().defaultClockClass()`.

Change-Id: I2923f3882dfd4dde31057962a1dd41f7b9edbd9a
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11965
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: add `streamClassDefaultClockClass()` methods

Add `streamClassDefaultClockClass()` methods on all relevant message
types, calling the
`bt_message_*_borrow_stream_class_default_clock_class_const` library
functions.

Change-Id: Iee1636048838d263b4407c60f8760870bd4ddb01
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11964
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

tests: improve debuggability of `cli/test-trace-copy.sh`

When something fails in `cli/test-trace-copy.sh`, we don't really know
why, as everything is sent to `/dev/null`.

Make the test record stderr and print it (using tap's `diag`) when
something fails.

Add the `diag_file` helper to make that easy.

Change-Id: I2eb85cd5fffb136cddf2457089681ee57c21e4f7
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11963
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: use `ComparableWithCStringView` on `operator!=` meant for `CStringView`

I don't recall exactly when I needed this, but it occured to me once
that the `operator!=` we provide for `CStringView` was getting picked up
for some unrelated type.  It just seems dangerous to provide a generic
templated operator like that that can match literally anything.

Use the same `ComparableWithCStringView` conditions here that we use for
`operator==` just above, so that it only gets used for what it was
intended for.

Change-Id: I798f1c8195c541f6188499dcfabfe1233a95a492
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11962
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: add static_assert in `CommonClockClass::uuid(const std::uint8_t *)`

Change-Id: I66399af51ab21c52e8fbc61cc61cae1b94f1e2e7
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11956
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

tests/lib/conds: test clock compatibility post conditions

Add tests for the clock class compatibility post conditions in
`iterator.c`.  There is one test for each `BT_ASSERT_POST_DEV` in
`assert_post_dev_clock_classes_are_compatible_one`.

`addClockClsCompatTriggers()` adds a bunch of triggers that make a
message iterator return two messages with incompatible clock classes.
The message types that play a role in the clock compatibility checks are
"stream beginning" and "message iterator inactivity".
`addClockClsCompatTriggers()` adds all possible combinations of using
these two message types for the first and second message.  Since it's
not possible to send a message iterator inactivity message without a
clock class, some cases are skipped.

The change in `Makefile.am` is needed for the test to work in builds
with `--enable-built-in-plugins`, for the `RunIn` utility to be able to
find the `utils` plugin.  This was not needed before, because all the
existing triggers make the program stop before `RunIn` attempts to load
the `utils` plugin.

Change-Id: Ib6c22923e4e9f02de64a87efddaee3c23d3891ea
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11811
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

Fix: lib/graph/iterator: consider clock classes from message iterator inactivity messages for clock correlation validation

Make the clock correlation validation code consider clock classes from
message iterator inactivity messages.  This is equivalent to what the
muxer component class does (see `MsgIter::_validateMsgClkCls()`).

Change-Id: I172d0015eb45508da7c412172b24aa4b0236ae9f
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12120
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2s: control span-lite's contract check based on BT_DEBUG_MODE

Enable or disable span-lite's contract checking feature based on
BT_DEBUG_MODE.  When contract checking is enabled and a contract
assertion fails, span-lite calls `std::terminate()`.

Change-Id: Ieba9a08712823daf4b77f135155cd15ec97900f5
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12119
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

tests/lib/conds: make `condMain()` take a span for args

Make `argv` a span, instead of having `argc` + `argv`.  There is no
point in this change except that it seemed possible and I wanted to do
it.

Change-Id: I1e7e76638e9e6365df05bb5c7b175d44b6badf6f
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12118
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: enable span-lite's `make_span()` function

Enable span-lite's `make_span()` function, which exists to compensate
for the lack of class template argument deduction in C++ pre 17.  This
allows creating spans without being too verbose.  This is not a standard
function, but the day we switch to C++ 17 and want to use `std::span`,
we can just replace all uses of `make_span()` with calls to
`std::span`'s constructor.

Make one version of `nonstd::make_span` (expected to be used in an
upcoming patch) available as `bt2c::makeSpan()`.

Change-Id: Ie28c1884d2c03c38a4b839096afae887fa54cc04
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12117
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

tests/lib/conds: store triggers in a vector

An upcoming patch needs to dynamically create and append triggers to the
trigger list.  Make the container a vector.  For simplicity, update
`condMain` to take a reference to a vector of unique pointers of
triggers (rather than inventing some abstraction).

Change-Id: Ic5e2fe6da151434ec515c041d4e1096193d29221
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12115
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

tests/lib/conds: change `nameSuffix` type to `bt2c::CStringView`

I'm not sure why `nameSuffix` is passed as a `const
bt2s::optional<std::string>&` currently.  `bt2c::CStringView` seems like
a good fit for it.  It avoids unnecessarily creating `std::string`s when
the caller passes a literal string.

Change-Id: I1412b77223fe01a70dda4c23558983f84dfb07d4
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12114
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

lib/graph/iterator.c: use `BT_IF_DEV_MODE` at a few places

Use `BT_IF_DEV_MODE` where it can easily replace an `#ifdef
BT_DEV_MODE`.

Change-Id: I452f200e09b71a1bd3719262bb6e10009b82d3de
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11968
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

lib: split "message-clock-classes-are-compatible" post-condition in multiple post-conditions

Currently, `clock_classes_are_compatible()` returns a bool indicating if
the clock classes referenced in the current batch of messages are
compatible with the clock classes seen previously by the iterator.  If
there is any incompatibility, whatever the reason, the
"message-clock-classes-are-compatible" post-condition check is
triggered.

A subsequent patch will want to test each individual clock class
incompatibility case, which is not easy if they all produce a message
with the same condition id.

Change `clock_classes_are_compatible_one()` to directly do the
post-condition checks, and give a unique name to each check.

Rename to `assert_post_dev_clock_classes_are_compatible_one`, in line
with the existing `assert_post_dev_expected_packet()` & co.

Adjust and rename `clock_classes_are_compatible()` accordingly.

Change-Id: I38937422f731c0cfebad70f183be97144d43e121
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11954
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: add `Uuid::Uuid(bt2c::CStringView)`

Add a `Uuid` constructor that takes `bt2c::CStringView` and remove those
that take a `const char *` and `const std::string &`.  The implicit
conversions to `bt2c::CStringView` should be able to handle all those
types.

Change-Id: I1714fd2aa38c18a047b097908af7b52afebeab72
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11953
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: make `CommonClockClass::uuid` take a `bt2c::UuidView`

This, combined with a following patch that makes the constructors of
`bt2c::UuidView` implicit, makes it possible to pass a `bt2c::Uuid`
object directly to `CommonClockClass::uuid()`.

Change-Id: I336763e128bef5de8e208bc7fafcd9df59d7672a
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11952
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: make setters return `*this`

Change setters (and other methods currently returning `void`) to return
`*this`.

This allows writing more compact code and avoid intermediate variables,
like:

    return graph->addComponent(
        *dummySinkCompCls, "the-sink",
        bt2::MapValue::create()->insert("key1", "val1").insert("key2", "val2"));

It doesn't always produce super readable code, but it's nice to have the
option.

In sub-classes of `BorrowedObject`, the methods return by value.
`ConstMessageArray::append` returns by reference.

Change-Id: I5af80c96b4c947dc217f4dff4b0bcb31885005e8
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11951
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: move `SelfSpecificComponent` to `internal` namespace

Change-Id: Icc3536de99be1e0de165e5ec990cffe73cf5be77
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11961
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: check return value in `CommonVariantField::selectedOption()`

Instead of ignoring the return value of
bt_field_variant_select_option_by_index(), check it with
BT_ASSERT_DBG().  This probably won't ever make a difference since
bt_field_variant_select_option_by_index() can't fail for now, but it
looks better.

Change-Id: I2354adfa1358ef0bd77b1d7c6604eef8ccaaa018
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11950
Tested-by: jenkins <jenkins@lttng.org>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: add message creation methods to `UserMessageIterator`

These are meant to be used by sub-classes of `UserMessageIterator`.

Change-Id: I4dbbd4a121218f432b322e8909b19f4a367f650b
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11847
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2c: change some `static inline` functions to just `inline`

I believe that using `inline` for functions in header is preferable to
`static inline`.  With `static inline`, the symbols resulting from the
functions are internal to each translation unit.  We may end up with
multiple copies of the same function in the final executable.  With
`inline`, the linker de-duplicates them.  This is better for code size
and cache and all that.  It may not be relevant for these functions, but
it would be good to set the precendent for the functions we will add in
the future.

The caveat with that is that all copies a function must be identical
(ODR), but I don't think we have a problem with that.

Change-Id: I62eeacf4e8326d70cb94aec33d6684453e1b8912
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11843
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>
Tested-by: jenkins <jenkins@lttng.org>

cpp-common/bt2: remove extraneous `protected`

Change-Id: I0771ac211b32f7cb5d07e87338efb22ba797a036
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11846
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: specify default constructor for `UnknownObject`

The constructor is trivial, specify a default constructor.

Change-Id: I9b9ecde9205e8dc57f7fe2f1da494db1c31633c2
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11845
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: use `std::unique_ptr` to manage `UserMessageIterator::_mSavedLibError`

clang-tidy points out that `UserMessageIterator` has a user-defined
destructor, but default copy constructors and assignment operators.
This is indeed dangerous: if we have an error in `_mSavedLibError` and
the `UserMessageIterator` object gets copied, the error will be released
multiple times, causing a use-after-free.

The most obvious fix would be to explicitly delete the default
definitions of the copy constructors and assignment operators.  But
I think it's a bit more idiomatic C++ to use some RAII type to manage
the error.  In this case `std::unique_ptr` is a good fit.  This lets us
delete the user-defined destructor.  The copy constructors and
assignment operators are implicitly deleted, since `unique_ptr` doesn't
have them.

Change-Id: Ia78bbda6a342192cb35c1b6963d3ad7fccaa7c3f
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11844
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

cpp-common/bt2: remove redundant assertions

Remove some redundant BT_ASSERT and BT_ASSERT_DBG.

 - Most `as*()` methods call a constructor that does the exact same
   check.  Some exceptions are field class types that have signed and
   unsigned variants, where the base class' assertion does not check the
   signedness.  Another one is `CommonNullValue::asNull`, where the
   `CommonNullValue` constructor does not take a lib object pointer, so
   does not do a check.
 - Things that end up constructing a `BorrowedObject` don't need to do
   a null pointer check, as `BorrowedObject`'s constructor does one.
 - `BorrowedObjectIterator::operator*` does not need an assertion on the
   length, as `operator[]` of the container is expected to do it (either
   directly, or rely on the preconditions in the lib).

Change-Id: I6f11680628607eee11164a2e9d3f9de7fc7468cb
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12052
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

Fix: cpp-common/bt2: more specific assertion in `CommonVariantWithIntegerSelectorFieldClass` constructor

It seems like we can check that the field class is not only a variant,
but more specifically a variant with integer selector.

Change-Id: Ib062779d957ba1e859247c891027121479ae3adb
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12051
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

Fix: cpp-common/bt2: don't pass a lib obj ptr when constructing a `CommonNullValue`

When trying to use `CommonValue::asNull()`, we get:

    /home/smarchi/src/babeltrace/src/cpp-common/bt2/value.hpp:1423:55: error: no matching function for call to 'bt2::CommonNullValue<const bt_value>::CommonNullValue(<brace-enclosed initializer list>)'
     1423 |     return CommonNullValue<LibObjT> {this->libObjPtr()};
          |                                                       ^

The `CommonNullValue` constructor takes no parameter, so don't try to
pass one.

Change-Id: I653ab947fc36134c8767f39185a9a2cd85511b90
Reviewed-on: https://review.lttng.org/c/babeltrace/+/12050
CI-Build: Simon Marchi <simon.marchi@efficios.com>
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>

tests/lib/utils: add RunIn::onMsgIterNext()

Add a virtual method on `RunIn` to execute some code in the context of
a message iterator's `next` method.

Change-Id: I6f8ca7a9596fd998f07c401c1859177c728a0b53
Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Reviewed-on: https://review.lttng.org/c/babeltrace/+/11834
Reviewed-by: Philippe Proulx <eeppeliteloop@gmail.com>