Summary: @public
Get newest goodness from facebook-clang-plugins
that allows us to change type of type_ptr during deserialization
We are modifying clang_ast_types.ml that fcp/clang-ocaml expects to
exist - we provide our own implementations of:
t_ptr, pointer_to_type_ptr, type_ptr_to_pointer
Reviewed By: @dulmarod
Differential Revision: D2498623
Summary: @public
First diff to give better language information in the frontend.
This information is necessary to understand when 'self' is objc keyword,
when 'this' is C++ keyword and when they are not.
Reviewed By: @ddino, @dulmarod
Differential Revision: D2489252
Summary: @public
There was an inconsistency with the type of `stats.json` created by `inferlib.py` and InferAnalyze. This diff cleans up the thing and uses two different files to saves the statistics to clarify what gets created by the analysis, i.e. `proc_stats.json` and what gets created all the time `stats.json`.
Reviewed By: @sblackshear
Differential Revision: D2500517
Summary: @public
qual_type was the same as type_ptr, so it was removed.
This commit makes it compatible with facebook-clang-plugins
Commands ran:
codemod --extensions ml,mli 'qual_type' 'type_ptr'
codemod --extensions ml,mli 'qt' 'tp'
If there is any other popular name for qual_type, let me know
Reviewed By: @dulmarod
Differential Revision: D2498289
Summary: @public Infer previously did not work correctly when a function returns the result of a skip function:
```
retUndef() {
x = undefined();
return x;
}
derefUndef() {
y = retUndef();
y.doSomething(); // Symexec_memory_error here, prevents spec inference
}
```
The problem is that angelic mode did not know to add the return value of `retUndef()` to the footprint.
This diff fixes the problem by adding return values marked with the `Aundef` attribute to the footprint.
This is done lazily (e.g., a value only gets added to the footprint when you try to deref it).
Reviewed By: @jvillard
Differential Revision: D2444929
Summary: @public
We no longer rely on existence of qt_raw field.
Update facebook-clang-plugins so that this field
is not even exported. This should give us some perf wins
since qual_type is written in a lot of places
Reviewed By: @dulmarod
Differential Revision: D2489308
Summary:
Using the qualified names of fields to create mangled names.
This removes the need to search for the fields in the tenv when translating a field access.
That was only done to build the correct mangled name,
The rest of the changes are using the qualified names for ivars and property names in the property module,
which is needed to make it work. Basically a big chain of changes of using qualified names to make the code
compile.
Summary:
This diff aims at removing the occurrences of the types in string form so that we can remove them
from the ast which will save space.
There is one occurrence left regarding attributes that will be handled later by Andrzej.
We build a function to create a string out of a function type used for name mangling.
Summary:
frontend and backend assume that Sil.Struct
doesn't have methods nor inheritance. In order to plug C++
classes we probably need them to be Sil.Class
Summary:
C function names depend only on their name (and their file is they are static, and their type if we are in C++)
This is more in sync with the actual semantics of C function names than the previous implementation.
Summary:
`get_resource_or_undef` attribute is weird and was causing problems for me in another diff.
This diff refactors the attribute categories to make resource and undef separate.
Summary:
Harness generation was using the autogenerated harness file as the source file,
but placing the harness code in the cfg belonging do a different source file. Fix this confusion
by using the source file of the cfg containing the harness code throughout.
Summary:
Some build systems compile files both under the project root and under temporary
directories (e.g., /tmp), so there need not be a single project root. Just use absolute paths
in the case that we can't resolve a relative path from the project root.
Summary:
Added two annotations @TrueOnNull and @FalseOnNull to be used for boolean functions to specify what value is returned when the argument is null.
Added model for TextUtils.isEmpty, which corresponds to the annotation
@TrueOnNull
static boolean isEmpty(@Nullable java.lang.CharSequence s)
Summary:
Make infer run with clang 3.7.0. Small changes are due to differences
in output of clang
NOTE: this diff will require recompiling clang (it takes time)
Summary:
Pass inheritance information to the backend
It also changes some functions in cTypes_decl and we are using type and decl maps to resolve these types
Summary:
This is the second of 3 stack diffs to deal with replacing the parser of types.
This diff is about changes to translate record types, as well as class types and enum
types. For class types and enum types we store the declaration pointer in the map of
types to find the type easier later.
For record declarations, we change the way we build record names.
Moreover, we don't translate typedefs anymore, because when we have a pointer to a typedef,
we can find the actual type it points to.
Summary:
This is the second of 3 stack diffs to deal with replacing the parser of types.
This diff is about general changes to the frontend to make it cope with the change. There
are two main challenges:
1. We create pieces of ast in ast_expressions, such as getters and setters. For that we create
custom types.
2. We store types in cMethod_signature for parameters and return type of functions. This was
stored as strings, but that means losing the pointer information which is vital to get the
sil types.
So this diff consists mostly of dealing with these challenges. It change the signature of
cMethod_signature and update modules accordingly.
To deal with the custom types, we build methods in ast_expressions for creating those types,
with a custom type pointer, like "internal_typeint". At the beginning of the translation we save
all these custom types in the map from type pointers to sil types that we build as we compute the
types, so that they are available later.
Another custom type that we build is a type of classes or pointer of classes based on the current
class. I found a simple way to deal with it, giving it a pointer "class_name", and then we know
how to translate those. Something I tried is to save the declaration of the current class and pass
that declaration around, but somehow that lead to pref regression, so I removed it in favor of this
more lightweight version.
Summary:
This is the first of 3 stack diffs to replace the parser of types in the clang frontend.
In this diff we remove the parser and the lexer and add a new module that does the
translation from ast types to sil types.
It is still incomplete, i.e. many c++ types are still not treated. However, all the
types that we were previously treating in C and ObjC are treated and some C++ ones, such
that the tests pass and we get good results in the apps.
Sometimes one needs to translate a record type when we havent translated the record itself,
so the translation of types and of records needs to be mutual recursive. I managed however
to get them into different modules and achieve the mutual recursion using higher order functions.
Summary:
System.getProperty can return null when the property is not found, and expects a non-null argument.
Add models for Infer and Eradicate to reflect that.
Summary:
each procedure has a different scope, so we can restart the fresh name generator and have more stable instructions in the cfg, that don't change when other procedures are changed
Summary:
Move proc_attributes to a separate module.
Field err_log, in common between proc desc and summary, can now be moved to ProcAttributes without creating cycles of dependencies.
Summary:
Create infer/src/backend/version.ml even if git is not installed or returns
errors. This is useful for source releases of Infer, which need to be able to
produce a sensible version number without a .git repository.
See also #167.
Summary:
There's a lot of overlap between the representation of a proc desc and a spec summary. This diff moves all the data in common to the single record proc_attributes defined in Sil.
This gives a unified way of accessing most of the data carried by a procedure, whether it is contained in a proc desc or a spec. Also, it ensures that there is a single flow of information from proc desc to spec in the back-end, making sure that the information represented stays consistent.
Summary:
Errors arising from overriding methods defined in other files were not reported, because during parallel analysis the clusters did not have access to overridden methods, so could not load their annotation.
Changed cluster generation to add location information for the methods overridden by the procedures defined in the current cluster.
Summary:
When someone runs --changed-only mode, there is a risk of corrupting the results
for future analyses. The problem is that changed-only mode does not analyze the callers of changed
procedures. If a subsequent analysis relies on the specs of one of these callers, they will be stale
and may give the wrong results. To be concrete, let's say we know `Parent.foo()` calls `Child.bar()` and we do the following rounds of analysis:
Analysis round 1: Analyze all files, including `Parent` and `Child`
Analysis round 2: Analyze `Child.bar()` only with `--changed-only flag`. `Parent.foo()` is now stale.
Analysis round 3: Add procedure `Parent.baz()` that calls `Parent.foo()`, analyze in (any) incremental mode.
The analysis will only analyze `Parent.baz()`. However, the specs for `Parent.foo()` are stale and may give us bad results for `Parent.baz()`. We want the analysis to re-analyze `Parent.baz()`, but before this diff it will not.
This diff fixes this problem by adding a `STALE` status bit to procedure summaries. In `--changed-only` mode,
the callers of a changed procedures are not re-analyzed, but their summaries are marked as stale. For both
`--changed-only` and regular incremental mode, callees of changed procedures that are marked as stale are
re-analyzed even if they have not changed. This is better than a more obvious solution like deleting stale
procedure summaries, since that would force the next analysis to re-analyze all stale procedures even if it
does not need the results for whatever analysis it is doing. This scheme implemented in this diff ensures
that each analysis only does the work that it needs to compute reliable results for its changed procedures.
Summary:
Use the new clang plugin that outputs biniou instead of Yojson. This binary format is more compact, which makes the frontend a little faster (5 to 10%).
@update-submodule: facebook-clang-plugins
Summary:
Passing the list of SuppressWarnings annotations detected during the compilation to InferPrint. The next step will be to add support for error filtering in .inferconfig and use the same mechanism. The annotation processor will generate an .inferconfig like config file and use it to suppress the reports.
Summary:
Procdesc comparison can be fragile because internal variable names
and source positions in a procedure can vary even if the procedure stays exactly the
same. This diff makes pdesc comparisons less fragile by defining structural comparsions
over instructions, nodes and expressions. These structural comparsions work by lazily
creating a mapping between names in the two procdesc's that and checking that the mapped
names are used consistently.
Summary:
Use the map of pointers to find method declarations and build method signatures.
Remove the need for having an extra map for method signatures (and remove that map).
Summary: Another step toward getting procedure-level incrementality for Infer. Here, we just assume that *all* procs in a file have changed when the file changes, but we will filter this list using the pd_changed field of the procdesc for each procedure in the future.
Summary: Handler.postDelayed keeps a persistent reference to its Runnable argument that may cause a memory leak if an Activity is reachable from the Runnable.
Summary: The Nullable checker reported FP's when a Nullable field/param was reassigned to a non-Nullable value in the footprint. This diff fixes the problem.
Summary:
The @NonNull annotation, with camel case, can now be used to inform Eradicate that some fields that are not initialized by the constructor can be initialized by other means, e.g. via dependency injection.
Summary:
see title. It will make number of arguments
less ridiculous and make it easier to share C/C++ structs.
Another diff that adds base class information for C++ will
follow.
This change is big enough to deserve separate diff
Summary:
This diff allows to use to make Infer failing when using the function `failwith`. This is especially useful to use while applying complex refactoring.
Summary: Infer cannot tell if a procdesc has changed across procedure runs. If we want procedure-level incrementality, it has to know how to compute this information. This diff implements this capability by comparing a procdesc to an existing one before it is saved to disk, and marking the new one as unchanged if applicable.
Summary:
Handle C++ method declarations and create cfgs for them.
Doesn't do:
Method calls (CXXMethodCall)
Using `this` expression in methods (including implicit ones)
Summary:
Refactor of creating method signatures. First step to use the map of pointers to find method declarations.
The idea is to have a function that creates a method signature from a declaration, later we can get the declaration from a pointer
and use this function to retrieve the method signature.
(authored by @dulmarod)
Summary:
Add basic translation for C++ `new` keyword.
Currently, it's modeled as simple `malloc` call.
Following constructs are still not working properly:
- array new `new [size_expr]`
- run initializer attached to `new` (such as `new int(5)`)
- `delete[]`
Summary:
When detecting a resource leak, Infer used to raise an Leak exception and then prevent the specs to be computed for the paths containing a leak. This diff prevents resource leak to stop the analysis.
Summary:
1. Unify the code now that MemberExpr has more information available and it can be shared with ObjCIvarRefExpr
2. Use type from decl_ref instead of expression type. For methods, expression type is useless (`<bound something something`>). For fields it should be the same
Summary:
Creating a persistent reference to an Activity leads to a nasty form of memory leaks (see http://android-developers.blogspot.com/2009/01/avoiding-memory-leaks.html, https://corner.squareup.com/2015/05/leak-canary.html). There are many ways to create a bad persistent reference to an Activity, but the most obvious one is via a static field.
This diff implements a very simple form of Activity leak checking by inspecting postconditions to see if a subtype of Activity is reachable from a static field (and it reports an error if so). This is a very simple and limited form of leak checking that does not understand the Android lifecycle at all. In particular, if one creates a persistent reference to an Activity and then nulls it out in `onDestroy` (a reasonably common pattern), this approach will wrongly report a bug.
Summary:
The methods in objc can have the same name in the same class, but one be instance and the other class,
so that we need to take the instance flag into account when defining unique names for ObjC methods.
Summary:
This is mostly useful to authors of annotation processors and
`javac`-based static analysis tools. The nullable return models
cover these packages pretty comprehensively (with the exception
of the various visitor classes). The non-nullable parameter models
are mostly there to help make the nullable return models more useful,
as some codepaths will pass a nullable object to one of the utility
classes but never actually invoke a method on the nullable object itself.
Summary:
While `-results_dir` is still the main place to look for specs files and to write reports,
it's necessary to load specs from multiple folders because some build tools that run Infer with a target-level granularity may need to move specs files around in order to get complete reports, whereas with this change they just need to keep track of the `specs` folders generated for each target, and pass them through `-lib dir1 -lib dir2 ... -lib dirN`
Summary:
The symbolic execution was not stopping in case an unitialized dangling pointer was
passed to a function and then dereferenced inside the callee.
What would happen is that a wrong footprint would be added to the unititialized pointer
at the end of the function call in the caller proposition.
This checks that if we do:
frame * new_footprint
checks that we do not add heap predicates to the frame into uninitialized local variables.
If we can identify the variable then we raise a danglind pointer dereference. If instead
we cannot give a good explanation we give an internal error.
The latter case should be temporary. We should find a general way to raise dangling pointer
deref instead of the internal error.
I also fixed the model of getc that was the way I found the problem.
Summary:
Update fcp version and make infer build with it.
It's not using new features yet, diffs will follow.
New stuff:
Proper type information in form of pointer->type
Expose more information about cxx classes (superclass info)
Add pointers to objc method decls when possible
Summary:
This is small code cleanup of the code to report leaks. No functional changes, just removing code that is longer used and reorganising the control flow.
Summary:
This commit is the result of
`find infer/src -name '*.ml' -or -name '*.mli' -exec ocp-indent -i \{\} \;`
and
`INFER_CHECK_COPYRIGHT=1 InferPrint`
Summary:
The old scheme for pruning away garbage from abducted retvars/abducted params passed by ref failed to eliminate garbage in the pure constraints (pi). This occasionally caused PRECONDITION_NOT_FOUND errors that stop the analysis.
Summary:
In preparation for C++ methods, we need to have type with
class, method, mangled (for overloading?)
1. Change objc method to support it
2. Do some renames to be less confusing
Summary:
Added phase to construct (incomplete) CFG from existing AST.
Individual instructions are not yet translated so any nonempty functions will
result in a runtime exception.
Added some very short examples of LLVM programs.
Summary:
This adds a sentinel check every time a function carrying a sentinel attribute
is called, regardless of whether we have a definition for that function or not.
Summary:
In objC we already prefix field names with classes.
It's better to make it consistent since it'll allow
us to share more code between C++ and objC
Summary:
Bumping up the version of the libraries javalib and sawja to 2.3.1 and 1.5.1 respectively in order to use the upstream fix about classnames with $$.
Summary:
This adds support for having lookup from pointer to decl in C frontend
NOTE: with this diff atdgen 1.6 or later is required.
To upgrade atdgen, please run `opam install atdgen.1.6.0`
Summary:
This makes infer C frontend compatible
with new scheme for naming which has:
1. plain name (like 'fun')
2. qualified name (reversed list like ['fun', 'class', 'top_class', 'namespace'])
Summary:
This attribute was used to tag arguments to variadic functions, as a way to
detect premature sentinels. The approach to detect premature sentinels has
changed making it obsolete.
Summary:
Treat `arrayWithObjects` as a special case of a sentinel attribute check. This
will make it easier to extend to other variadic functions that use a sentinel
attribute.
This also removes the need for the `Sil.Avariadic_function_argument` attribute,
which will be removed in a subsequent diff.
Summary:
the name of the return variable of a procedure only depends on the name of that procedure. This simplifies the need for the procedure description in a couple of places
Summary:
This is a refactoring diff with no functional changes in order to move the code using the procedure description of the callees in a single place, in order to replace the approach to use a summary instead in a subsequent step.
Summary:
This reverts commit 306f5b71c24042c89f71848898402cbc9269c543.
Turns out that developers think that this bugs should be fixed. So leaving it in for now until I gather more information.
Summary:
@public
There are many FPs of the form init method that contains
if ((self = [super initWithFrame:frame])) {
...
}
return self;
then an object being initialised with that constructor and added to an array or dictionary.
There we flag NPE and very likely that won't be a bug. So I'm removing the option for self
to be nil in the constructor, which should solve the problem.
Test Plan: Changed the relevant test.
Summary:
@public
Previously, if the close() method was throwing an exception, then code overriding the file attribute with a mem attribute would be skipped, resulting in reporting a wrong resource leak. This diff fixes this.
Test Plan: Added new end-to-end tests which would previously have been failing
Summary:
@public
Add support for default function arguments.
As a side change - always create cmethod_signature for a function
Test Plan:
1. Call function with default parameter and confirm that it gets parsed and reports null dereference (B5 but still). It didn't before.
2. Created a test case
Summary:
@public
This is a non-functional refactoring to remove the need for having the procedure description of the callee when execution function calls.
Test Plan: Infer CI. Expecting no change in the results.
Summary: @publicThe first argument of builtin calls in C gets translated twice, which is bad if the argument is a side-effecting expression like a function call.
Test Plan: Attached test previously reported a memory leak because the translation introduces an extra call to malloc(), now reports nothing.
Summary:
@public
Translate CXXStaticCastExpr
Test Plan:
Add test, confirm that it gets translated.
Also create example to see that infer reports null dereference with this change:
struct X { int a; };
int main() {
X *x = static_cast<X*>(nullptr); // <- reports now
//X *x = (X*)nullptr; // <- reported before
return x->a;
}
Summary:
@public
Currently InferAnalyze always adds bucket to the message. Later, python code
strips it, but not everywhere. Changes:
1. Since it's easy to not write bucket in ocaml, stop writing them by default.
2. Add option to print them to InferAnalyze and pass it if infer is in debug mode.
Test Plan:
1. Run on openssl, confirm that no bucket info is written to stdout and csv
2. Run on small example in debug mode and see buckets on stdout
Summary:
@public
Make c frontend understand CXXNullPtrLiteralExpr.
Note that the implementation differs from GNUNullExpr
Test Plan:
Create function that returns nullptr:
int* getPtr() {return nullptr;}
look at specs:
InferPrint infer-out/specs/getPtr\{831F\}.specs
Procedure: getPtr
int *getPtr()
Timestamp: 1
Status: INACTIVE
Phase: RE_EXECUTION
Dependency_map:
TIME:0.002853 s TIMEOUT:N SYMOPS:10 CALLS:1,0
ERRORS:
--------------------------- 1 of 1 [nvisited: 1] ---------------------------
PRE:
POST 1 of 1:
return = null:
----------------------------------------------------------------
Add test for it
Summary:
@public
Using InferBuiltins.assume previously caused an assertion failure in the analyzer. Fixed this, and fixed the implementation of the assume builtin to block when the assumed condition cannot hold.
Test Plan: Added several new tests.
Summary:
@public
Sorting the fields in structs and classes. Was needed in the backend and forgotten.
Fixes the github issue https://github.com/facebook/infer/issues/90.
Test Plan: Added a new test that shows that we now get a spec for the example from the github issue.
Summary:
@public
The empty string '' is a valid classpath information. This diff parses it and skips it when detecting the parts of the classpath. This case is happening when analyzing Buck. This one be one possibility why Infer does not load any bytecode in some cases, leading the
TODO: print error message
failure happening with the release (now `Failed to load any Java source code`).
Test Plan: Infer CI.
Summary:
@public
This adds basic support for function attributes in Sil, and for translating
attributes from the clang frontend to these new Sil attributes. For now only
the sentinel attribute is translated.
Note that attributes normally have parameters, but they are currently missing
from the clang plugin.
Test Plan:
Add
(match Sil.get_sentinel_func_attribute_value (Cfg.Procdesc.get_attributes callee_pdesc).Sil.func_attributes with
| Some _ -> L.out "found sentinel attribute!\n"
| _ -> ());
between lines 947 and 948 of symbExec.ml, then analyze a file containing:
int add_all_ints(int a, ...) __attribute__ ((sentinel));
int foo(void) { return add_all_ints(1, 2, 3, (void *)0); }
then `grep 'found sentinel' infer-out/log/analyzer_out`
-> the sentinel attribute is correctly passed from the frontend to the backend.
Summary:
@public
- Remove `No such file or directory` unrelated errors when building Infer for Java only:
line 0: cd: infer/../facebook-clang-plugin: No such file or directory
- Remove makefile comments from stdout
Test Plan:
Ran this command without any checkout of `facebook-clang-plugins`
make -C infer java
The misleading error messages are gone.
Summary:
@public
The clang location information is described in an incremental way: each location information is a delta with respect to the previous one in the AST. This is based on a the visit of the AST nodes which corresponds to the order in which the lines are printed with the standard clang AST dump:
clang -cc1 -ast-dump filename.c
This diff adds a preprocessing phase to the front-end so that location information is composed during a visit, and explicit location information is used instead.
In the case of include files, we report the last known location before including the file.
The current file for a function is the file where it is defined. So if a function is entirely defined in a .h file, then the location information will consistently be about the .h file. If instead a function is defined in the source file being analyzed, and some AST nodes come from macro expansion, line information will refer to the original file.
The front-end tests reveal that the location information was incorrect in a few dot files.
Test Plan: arc unit, after having fixed the wrong location in the existing .dot files
Summary:
@public
This changes "Starting analysis" into
"Starting analysis (Infer version XXX)".
Test Plan:
infer -- clang -c hello.c
shows "Starting analysis (Infer version git-6b9fb8838bcabd2af881554d296963a849b14f50)"
Andrzej Kotulski
Cristiano Calcagno
Dulma Rodriguez
jrm
Sam Blackshear
martinoluca
Irene Papakonstantinou
Jules Villard
Rohan Jacob-Rao
Jonathan Keljo
Dino Distefano
Open Source Synchronization