typename(table()) apparently always resulted in a set[] type
being rendered. Make the yield type of an unspecified table
ANY so that type->IsSet() ends up false.
While at it, also render unspecified types as table(), set() and
vector() rather than vector of void, set[] or table[] of any which
IMO should help to figure out what's going.
Really, they both should be count. But, they were getting provided as an
integer. Port is easy since it is backed by an unsigned value. Enums
*should* be unsigned, but aren't. This doesn't address that, it just
takes the other name for this operator (absolute value) and makes the
enum value positive if it's negative.
This fixes a case where using the size of operator on enum/port values
in certain contexts (like the default parameter of a struct) would cause
an internal error.
* origin/topic/vern/zam-regularization: (33 commits)
simpler and more robust identification of function parameters for AST profiling
fixes to limit AST traversal in the face of recursive types
address some script optimization compiler warnings under Linux
fix for -O C++ construction of variable names that use multiple module namespaces
fix for script optimization of "opaque" values that are run-time constants
fix for script optimization of nested switch statements
script optimization fix for complex "in" expressions in conditionals
updates to typos allow-list reflecting ZAM regularization changes
BTest updates for ZAM regularization changes
convert new ZAM operations to use typed operands
complete migration of ZAM to use only public ZVal methods
"-O validate-ZAM" option to validate generated ZAM instructions
internal option to suppress control-flow optimization
exposing some functionality for greater flexibility in structuring run-time execution
rework ZAM compilation of type switches to leverage value switches
add tracking of control flow information
factoring of ZAM operation specifications into separate files
updates to ZAM operations / gen-zam regularization, other than the operations themselves
type-checking fix for vector-of-string operations
ZVal constructor for booleans
...
* origin/topic/vern/script-opt-maint.Aug24:
minor optimization of boolean comparisons
fix & regression test for GH-3839 (spurious warnings for "when" constructs)
This needed a small tweak in the deserialization, since each roundtrip
would otherwise pad the prior pattern with an extra /^?(...)$?/.
This expands the language.set test to also verify serializing/unserializing for
sets, similarly to tables in the previous commit.
This allows additional data roundtripping through JSON since to_json() already
supports tables. There are some subtleties around the formatting of strings in
JSON object keys, for which this adds a bit of helper infrastructure.
This also expands the language.table test to verify the roundtrips, and adapts
bif.from_json to include a table in the test record.
@vpax reported surprising behavior when working with "void values".
While these are not exposed to script land, plumb the places he
pointed out are causing confusing behavior.
Closes#3640.
Since 81a9745fb3, the assert condition is
evaluated twice. This leads to unexpected behavior when cond has a side
effect like publishing a message or creating a log stream or filter.
Found while using the following in ad-hoc testing code and wondering
why two messages were published.
assert publish(Cluster::worker_topic, hello, "abc")
Provide a script accessible way to introspect the DFA stats that can be
leveraged to gather runtime statistics of the underlying DFA. This
re-uses the existing MatcherStats used by ``get_matcher_stats()``.
Not sure how useful this is (and the implementation isn't optimized in
any way), but seems reasonable for consistency.
Vern suggested that set[pattern] can already be achieved via
set_to_regex(), so left out any set[pattern] variants.
A bit larger follow-up to what Tim pointed out: Function prototype descriptions
previously used semicolons to separate parameters.
Switch to use commas when a RecordType is used as function parameter.
Use existing "func_args" naming for consistency.
* origin/topic/vern/script-opt-maint.Sep23:
fix for ZAM statement-level profiling (broken by GH-3199)
ZAM fixes for compatibility with GH-3249 changes
-O gen-C++ fixes for compatibility with GH-3249 changes minor -O gen-C++ BTest updates
minor BTest reordering to diminish differences with script optimization
Initializing fields of recovered records caused running &default expression
of fields just so that they are re-assigned in the next step with the
recovered fields. The second test case still shows that the loop var
is initialized as well even though that's not needed.
Add tests for iterating over records with &default attributes for both,
tables and vectors.
Fixes#3267
* origin/topic/timw/3059-set-vector-conversion:
Fix conversion with record types
Add conversion between set and vector using 'as' keyword
Add std::move for a couple of variables passed by value
This is based on the discussion in zeek/zeek#2668. Using &default with tables
can be confusing as the default value is not inserted. The following example
prints an empty table at the end even new Service records was instantiated.
type Service: record {
occurrences: count &default=0;
last_seen: time &default=network_time();
};
global services: table[string] of Service &default=Service();
event zeek_init()
{
services["http"]$occurrences += 1;
services["http"]$last_seen = network_time();
print services;
}
Changing above &default to &default_insert will insert the newly created
default value upon a missed lookup and act less surprising.
Other examples that caused confusion previously revolved around table of sets
or table of vectors and `add` or `+=` not working as expected.
tbl_of_vector["http"] += 1
add tbl_of_set["http"][1];
This marks every identifier used within an attribute as seeds. The scenario
this avoids is functions referenced through attributes on unused tables or
record types (&default, &expire_func, ...) being dinged as unused as
that's rather confusing.
Also adds test for the above and a light smoke test into language/ as it
doesn't appear we had coverage here.
Closes#3122
Using break in either of the hooks allows to suppress the default reporter
error message rather than suppressing solely based on the existence of an
assertion_failure() handler.
