Summary:
Add let*/+ syntax to `result` types to simplify all the applications of
`>>=`, `>>|` that are followed by a binding (eg `>>= fun x -> ...`) in
pulse.
Reviewed By: skcho
Differential Revision: D19940728
fbshipit-source-id: 4df159029
Summary:
We can already tell that a summary cannot be applied by raising
`Contradiction`, so use this mechanism to stop applying a summary if the
number of formals doesn't match the number of actuals provided.
Previously we would return an option type and `None` in case of
mismatch, on top of the `raise Contradiction` mechanism (used for
aliasing and arithmetic contradictions).
This changes the behaviour of pulse in this case: before we would skip
over the function call, but now we stop the analysis.
Reviewed By: dulmarod
Differential Revision: D19940729
fbshipit-source-id: 6def40cd6
Summary: Once we identify a weakSelf variable that is being used in a Noescape block, we want to report only the first occurrence.
Reviewed By: skcho
Differential Revision: D19941502
fbshipit-source-id: 2b6d4648b
Summary: For each variable that we identify as a captured strong self, we want to report only the first occurrence.
Reviewed By: skcho
Differential Revision: D19940031
fbshipit-source-id: f38f642c9
Summary: When we discovered that a strongSelf var was not checked for null, we then report in each occurrence which is spammy. Now we report only the first occurrence. To achieve that, we store a `reported` flag in the domain that gets set to true after we report once, and we only report if it's false.
Reviewed By: jvillard
Differential Revision: D19877218
fbshipit-source-id: c44109ae9
Summary:
This adds `let*/+`, `and*/+` operators for Option. See [the manual](https://caml.inria.fr/pub/docs/manual-ocaml/manual046.html) for more information.
Example usage:
```
let foo =
let open IOption.Let_syntax in
let* a = get_optional () in
let* b = get_another_optional () in
return (a + b)
```
Reviewed By: jvillard
Differential Revision: D19880033
fbshipit-source-id: c7998b0c6
Summary:
Previous implementation supported only stringy params (strings and
stringified bools). Current one exposes a proper variant `Annot.t`,
with support for all possible param values in Java except
numbers (more on that below).
This change is required for implementing `Nullsafe(LOCAL)` as the
annotation used to specify nullsafe behaviour has a more complex
structure than what we've dealt with before.
**Why support for number values was not added**: supporting numbers
requires using `int64`. Unfortunately, adding another variant `Vnum
int64` to `Annot.t` causes a runtime failure on assert in
`MaximumSharing.ml:133`. It seems that it might be enough to flip
`fail_on_nonstring` from `true` to `false`, but since this would
require additional testing and is not required for my case, I'll leave
checking this to whoever needs to use numeric annot params in future.
Reviewed By: ezgicicek
Differential Revision: D19855923
fbshipit-source-id: 878e33856
Summary: No reason to use a set when an integer will suffice. This further reduces GC churn.
Reviewed By: fgasperij
Differential Revision: D19888300
fbshipit-source-id: 9fc8c73f5
Summary: Queues are implemented using a circular array, so should be less GC-heavy than continually allocating/freeing list nodes.
Reviewed By: jberdine, fgasperij
Differential Revision: D18504104
fbshipit-source-id: 93d29c253
Summary:
Building the call graph should be done only in the scheduler process after having forked all workers. This was achieved by a lazy init pattern, whereby the first time `next` was called, it would build the call graph, on the assumption that `next` is only ever called in the scheduler after forking.
D19769741 made this compulsory regardless the scheduler by passing a thunk to `ProcessPool` which is called to obtain the actual scheduler, on the right process and after the fork.
This means we don't need the custom lazy init logic any more. In addition, that set up used a DB query to overapproximate the number of procedures to analyse, because this was supposed to be provided *before* forking. Now this is also not needed, and on top of that we can provide the exact number after building the call graph.
Reviewed By: ezgicicek
Differential Revision: D19833974
fbshipit-source-id: 7f6d51d93
Summary:
In Inferbo, the bottom memory is introduced when a node is unreachable by pruning, i.e.
`[[e]] <= [0,0]` on `prune(e)`. This diff distinguishes whether `[[e]]` is `[0,0]` (unreachable)
or bottom (it could not evaluate `e` by some unknown reasons).
Reviewed By: ezgicicek
Differential Revision: D19902046
fbshipit-source-id: 7706017d6
Summary:
More newer = more better.
This flips the Not_found -> Not_found_s switch, and forbids a bunch more
polymorphic comparisons (mostly turned into `int` comparisons for
convenience). Earlier diffs prepare for this so this diff is only about
breaking changes in the API, of which there are only a few.
Reviewed By: jberdine
Differential Revision: D19861583
fbshipit-source-id: fe54ce8f0
Summary:
Core v13 APIs stopped raising `Not_found` and instead raise
`Not_found_s`, which wreaks havoc in our codebase. Carefully inspect
each `Not_found` and add `Not_found_s` where needed (that way it's
compatible with both Core v12 and v13 for now).
Reviewed By: jberdine
Differential Revision: D19861585
fbshipit-source-id: 9a5361ae9
Summary:
The big one:
- stop using polymorphic `<>`, `<`, `>`, ..
- add `<>` to `PolyVariantEqual` escape hatch now that `<>` is as taboo as `=`
- Interestingly, there were a lot of uses of `Z.(x < y)`, which although
they seem to use `Z.lt` actually used polymorphic comparison. The actual
comparison infix operators of `Z` are cleverly hidden in `Z.Compare`
instead, which makes them impractical to use...
Reviewed By: jberdine
Differential Revision: D19861584
fbshipit-source-id: 5dce08ad9
Summary:
Polymorphic compare is bad, 'mkay? I also tried changing this to do what
the comment says the function does (keep the first term instead of the
"largest" term) but that resulted in test changes so I kept the original
behaviour instead.
Reviewed By: jberdine
Differential Revision: D19861587
fbshipit-source-id: d57fd4a02
Summary:
That trick briefly worked when it was introduced but I think a change in
merlin broke it again as I see "sexp_list", "sexp_option" in the types
shown by merlin again. Disgusting!
Remove useless hack. Also referring to types such as `sexp_list` is
deprecated in core v13 for some reason.
Reviewed By: jberdine
Differential Revision: D19861586
fbshipit-source-id: 1c4c3af13
Summary:
This adds a violation of baos.topl found in github/seata/seata. However,
it is not a bug (see comment in commit).
Reviewed By: ngorogiannis
Differential Revision: D19518641
fbshipit-source-id: e219245ee
Summary:
Since Javalib 3.2, a new feature allows to rewrite
methods that contain (some specific form of) closures. Infer
now uses it. When loading each class we rewrite them and
new classes generated by Javalib to implements closures
(i.e. Java interfaces)<
Reviewed By: ngorogiannis
Differential Revision: D19389227
fbshipit-source-id: 245dd4404
Summary:
When finding a proper constructor for `std::make_shared`, the given parameter types are sometimes
slightly different, e.g., const int vs int. This diff loosens the condition of the types on finding
constructors.
Reviewed By: ngorogiannis
Differential Revision: D19743198
fbshipit-source-id: f90213109
Summary: Instrument the ProcLocker to get the aggregated systime that is used by `lock` and `unlocks`.
Reviewed By: ngorogiannis
Differential Revision: D19814554
fbshipit-source-id: 5fd928b9c
Summary:
This was introduced in D19770219, and skcho caught it, but for some
reason I seemed to have a blind spot and confused the semantics of `instanceof`.
Reviewed By: skcho
Differential Revision: D19813031
fbshipit-source-id: d939b981b
Summary:
This diff fixes the clang translation for switch statement. It assumed that `default:` comes always
at last, which introduced some unreachable nodes inadvertently, e.g. when `default:` comes at first.
Reviewed By: dulmarod
Differential Revision: D19793138
fbshipit-source-id: 1e8b52c0d
Summary: After looking at some reports with blocks inside blocks, it seemed more obvious that adding which method we are talking about makes more clear which block we are talking about.
Reviewed By: mityal
Differential Revision: D19789285
fbshipit-source-id: 20e0e6804
Summary:
The domain has a notion of lock state (taken/not taken) and any access occurring will remember the current lock state.
Methods in Java can be `synchronized` meaning the lock is taken automatically at method start and release on return.
Currently, instead of starting analysis of a `synchronized` method with an initial lock state of "lock taken", the summary would be computed as if there is no lock, and then a caller would peek at whether the callee is `synchronized` and change the callee summary accordingly before taking it into account. Also, when an access happens (not a call) the analysis always consults the pdesc of the current method to check whether the method is `synchronized`.
Do things the right way: if method is `synchronized`, start with the lock taken. When the method exits, the lock-state postcondition is computed by releasing once the lock.
This is a behaviour-preserving change.
Reviewed By: mityal
Differential Revision: D19742559
fbshipit-source-id: 1d0fce3f6
Summary: This diff makes the taint analysis in Inferbo inter-procedural: adding symbolic taint values and substitute them at function call statements.
Reviewed By: ezgicicek
Differential Revision: D19411022
fbshipit-source-id: 4ff9a590a
Summary:
We already warn about lack of nullable annotations in `equals()`, and even have a specialized error message for that.
But lack of an annotation is not as severe as direct dereference: the
latter is a plain bug which is also a time bomb: it will lead to an NPE not immediately.
This is widespread enough to be reported separately.
Reviewed By: dulmarod
Differential Revision: D19719598
fbshipit-source-id: a535d43ea
Summary:
Since we fixed a bug in implementation of FalseOnNull (see stack below),
we can finally ship this change.
Side note: this change is essential for the follow up diff (which adds extra check
for user-defined implementations of equal()), without it the follow
up change would introduce a lot of false positives.
Reviewed By: ngorogiannis
Differential Revision: D19771057
fbshipit-source-id: 7d7cf1ef7
Summary:
If we managed to whitelist a function as TrueOnNull, we should teach
nullsafe the nullability of its arguments, otherwise it will ask not to
pass null here.
This fixes a silly FP warning, see the test.
Reviewed By: dulmarod
Differential Revision: D19770341
fbshipit-source-id: 0f861fae1
Summary:
Yay, the previous refactoring finally makes it possible to do some actual
changes to the code in `TypeCheck.ml`!
Changes in this diff:
1. Fixes the bug: TrueOnNull and FalseOnNull were working only for
static methods. Surpsingly nobody noticed that. It is because the first
argument for non-static method was `this`.
2. Behavior change: TrueOnNull/FalseOnNull were not working correctly
where there are several argumens. See the task attached for the example
of the legit usecase. Now the behavior is the following: if there are
several Nullable arguments infer nullability for all of them.
Reviewed By: skcho
Differential Revision: D19770219
fbshipit-source-id: 7dffe42cd
Summary:
This diff proceeds clean up the mess in TypeState.ml
This refactoring unblocks the change in the logic for true on null and
allows to fix a bug, see follow up diffs.
The current code is trying to do two things at once: processing boolean
results (which normally need manipulations with the argument of a
function + additional logic for containsKey), and comparisons witn null
(which requires manipulation with return value in typestate plus check
for redundancy).
All this was done in sort of generic fashion, which, among other, lead
to weird situations for edge cases, e.g. in processing containsKey we used to override nullability
already inferred as a non-null twice with losing information about
original nullability; which made logs weird).
Reviewed By: ngorogiannis
Differential Revision: D19768795
fbshipit-source-id: c928e2cff
Summary: The main job the schedulers do is building their work queues. That's being performed before the workers are forked which means they get copied into all of them. These changes push the initialization of the schedulers just after the forking takes place.
Reviewed By: ngorogiannis
Differential Revision: D19769741
fbshipit-source-id: 0b20ddd5c
Summary:
This diff is part of cleaning up of Typestate.ml mess to make it
somewhat maintainable.
This method is always called with `default` equivalent to (exp,
typestate); also there is no semantical need to return typestate because it never
gets changed.
Reviewed By: ngorogiannis
Differential Revision: D19767366
fbshipit-source-id: 173dcbbca
Summary:
This refactoring is made possible by previous stack, which ensured we
don't do two completely different things in one code anymore (processing
results of functions returning booleans and objects in generic fashion).
Follow up diffs will clean up the code for Prune(a != zero) case.
Reviewed By: dulmarod
Differential Revision: D19745050
fbshipit-source-id: 61d3d02ad
Summary:
This refactoring unblocks the changes in follow up diffs (plus fixes a
bug).
So what was happening?
Each comparison with null leads to CFG being splitted into two branches, one branch
is PRUNE(a == null) and another is PRUNE(a != null).
PRUNE(a != null) is where most of logic happens, it is the place where
we infer non-null nullability for a, and this is a natural place to
leave a check for redundancy.
Before this diff we effectively checked the same thing twice, and used
`true_branch` (only one of 2 instruction will have it set to true) as a symmetry breaker.
This diff removes the `true_branch` checks, but leaves only one call out
of two, hence breaking symmetry in a different way.
## Bug fix
The code around the removed check was (crazily) doing two things at
once: it processed results of (returning booleans!)
TrueOnNull-annotated functions AND
results of (returning Objects!) other functions, using the fact that all
of them are encoded as zero literals (sic!).
Not surprisingly that lead to a bug where we accidentally call the check
for non intended places (arguments of trueOnNull functions), which lead
to really weird FP.
This diff fixes it.
Reviewed By: dulmarod
Differential Revision: D19744604
fbshipit-source-id: fe4e65a8f
Summary:
These two methods are called in processing prune instructions, when
instruction is Prune(expr == null) and Prune(expr != null), to correctly
infer nullability in corresponding branches.
Typechecking underlying expr makes little sense for two reasons:
1. In practice, expr it is as simple as a temporary SIL variable
2. If the idea is defensively typecheck everything for case when SIL
produces crazy expressions, well, that is not going to work: the code
around ignores many other forms of expressions, e.g. everything where
expr = <something not equal to null literal>. So this is inconsistent.
This will simplify further cleanup, see follow up diffs
Reviewed By: ngorogiannis
Differential Revision: D19743826
fbshipit-source-id: 319a80ee7
Summary:
The whole TypeCheck.ml is exceptionally hard to read and maintain, lets
clean up it a bit.
Reviewed By: ngorogiannis
Differential Revision: D19743632
fbshipit-source-id: c24c21a85
Summary:
The goals are:
- Increase precision in C-languages by ditching access paths.
- Help with eventually sharing the abstract address module with RacerD.
- Reports are now language-mode specific (eg `->` in clang vs `.` in Java).
It's not exactly access expressions used here. Instead the pattern `(base, access list)` is used where `access` is `HilExp.Access.t`. This is done to ease the way `deriving` is used for creating two comparison functions, one that cares about the root variable and one that doesn't; and also because the main function that recurses over accesses (`normalise_access_list`) visits the accesses from innermost to outermost.
Also, kill some dead code.
Reviewed By: skcho
Differential Revision: D19741545
fbshipit-source-id: 013bf1a89
Summary: We don't use allocation costs in prod at the moment. There is no plan to do so in the near future. Let's not report them anymore and also save some space in `costs-report.json`.
Reviewed By: skcho
Differential Revision: D19766828
fbshipit-source-id: 06dffa61d
Summary:
This diff introduces two issue types: `BUFFER_OVERRUN_T1` and `INFERBO_ALLOC_IS_TAINTED`, which
denotes tainted values are used in array accesses and memory allocations, respectively.
Note that the taint analysis is intra-procedural for now.
Reviewed By: ezgicicek
Differential Revision: D19410536
fbshipit-source-id: af85148ec
Summary:
This diff adds a taint domain in Inferbo. The taint value will be used to find vulnerable array
accesses in the following diffs.
Reviewed By: ezgicicek
Differential Revision: D19391028
fbshipit-source-id: 566b4c0fe
Summary: Fixing bugs in the report_unchecked_strongself_issues_on_args function: we were not recursing if the first arg wasn't a var, and we were not removing the annotations from self in instance methods. Here we fix those issues.
Reviewed By: ngorogiannis
Differential Revision: D19662886
fbshipit-source-id: 03820961e
Summary:
The RestartSchedulerTests were failing because they are run in parallel by OUnit and all share the same output
directory. This makes the different tests collide since they are using the same file locks directory.
Reviewed By: jvillard
Differential Revision: D19765647
fbshipit-source-id: 5390ad14e
Summary:
This test tests PropagatesNullable and TrueOnNull/FalseOnNull
annotations.
Both tests suites grew big so it is hard to observe them at glance and
make changes.
I could not figure out better name for TrueFalseOnNull.java, it is sort
of silly but I optimized for searchability, "FalseOnNull" will be
directly searched and "TrueOnNull" will be searched in IDEs that are
smart enough.
Reviewed By: skcho
Differential Revision: D19724512
fbshipit-source-id: 703961342
Summary:
The problem with is_override that it can be misleading: it does not
check that that class of the first method is a subtype of the second; in
fact it totally ignores the classes.
This method is publicly exposed so lets just call what it does.
Reviewed By: jvillard
Differential Revision: D19724295
fbshipit-source-id: dc0919193
Summary: This diff returns non-symbolic value (top) for unknown external function calls because the symbolic values sometimes make it hard to understand costs.
Reviewed By: ezgicicek
Differential Revision: D18685715
fbshipit-source-id: 1b39c718b
Summary:
Pulse has an extra invalidation mechanism (introduced in D18726203) to prevent something invalid (e.g. `null`) to be passed by reference to an initialisation function. Therefore, it havocs formals passed by reference to skipped functions. However, I don't think this makes sense in Java. So, let's turn it off.
A nice consequence of this is that in impurity analysis, we do not consider functions that call skipped library calls with object arguments as writing to their formals.
Reviewed By: skcho
Differential Revision: D19697110
fbshipit-source-id: 6e3a71f2a
Summary:
To emulate the `ThreadSafe` contract in C++/ObjC, reporting was gated behind a check that ensured a C++/ObjC class has a `std::mutex` member (plus other filters). This is reasonable, but it has some drawbacks
- other locks may be used, and therefore must be added to the member check;
- locking mechanisms that use the object itself as a monitor cannot be modelled (`synchronized` in ObjC)
RacerD already has `ThreadsDomain` which models our guess on whether a method is expected to run in a concurrent context, and which in C++/ObjC boils down to whether the method non-transitively acquires a lock. This should be a good enough indicator that the class should be checked regardless of whether the locks are member fields. This diff gates the C++/ObjC check on that abstract property.
Reviewed By: dulmarod
Differential Revision: D19558355
fbshipit-source-id: 229d7ff82
Summary: This diff removes a dead field, `is_cpp_nothrow` and `is_cpp_noexcept_method`.
Reviewed By: jvillard
Differential Revision: D19489417
fbshipit-source-id: 971a7f533
Summary:
The ProcLocker uses files as locks and relies on the guarantees of the `Unix.open_file` function when using `O_CREAT` and `O_EXCL` simultaneously.
- `setup`: creates a directory for the lock files inside `infe-out` and deletes its content if it already existed.
- `clean`: does nothing for now. Any file locks that may have been left unlocked are removed by the `setup` in the next run. This way the user can see what locks were taken if the program crashes.
- `lock_exn`: try to lock the `Procname` and if it can't releases all the locks that is currently holding.
- `unlock`: removes the corresponding file.
Reviewed By: ngorogiannis
Differential Revision: D19639402
fbshipit-source-id: e02f277ff
Summary:
Revert incomplete/incorrect translation of `synchronized` in ObjC.
The current translation is incomplete because
```
syncrhonized(foo){
return;
}
```
should be translated as
```
__set_locked_attribute(foo);
__delete_locked_attribute(foo);
return;
__delete_locked_attribute(foo);
```
but instead we get
```
__set_locked_attribute(foo);
return;
__delete_locked_attribute(foo);
```
The same applies for `break`/`continue` etc
Reviewed By: skcho
Differential Revision: D19718882
fbshipit-source-id: fc49ef529
Summary:
- Thread the two types into one instead of having a record where the `path` field doesn't always make sense (`Class` case).
- Improved pretty printing of class objects (java only).
- Move starvation-specific stuff out of `AbstractAddress` (eg `make_java_synchronized`).
- Slight optimisation of `apply_subst` for when a parameter is used without additional accesses inside a method (then, the substitution need not modify the term substituted for the parameter in any way).
Reviewed By: ezgicicek
Differential Revision: D19639922
fbshipit-source-id: 1cebecf5d
Summary:
`String` and `StringBuilder` both implement `CharSequence`. Let's generalize the model for `String` to `CharSequence` wherever possible and add missing models for
- `StringBuilder.append`
- `StringBuilder.toString`
Reviewed By: skcho
Differential Revision: D19558009
fbshipit-source-id: 0dfdb21af
Summary:
Java's String models were broken for
- initializing a String object with a locally defined constant string (which is an `Object*` in SIL).
- initializing a String object with a `char`/`byte` array
This diff fixes them and also adds models for `new String ()`.
Reviewed By: skcho
Differential Revision: D19662180
fbshipit-source-id: 23968d0aa
Summary:
When the expression resolving to a function to be called could not be
translated to either a proc name or at least an access path, the HIL
translation code would crash. However, this is perfectly possible. Moreover, no
one actually uses the payload of the `Indirect` datatype so no complication
arises from generilising its type to `HilExp.t`, as done in this diff.
Reviewed By: ngorogiannis
Differential Revision: D19691127
fbshipit-source-id: 4c0400ab7
Summary:
`replace_make_shared` is finding a constructor by iterating all fields of type structs, which may be
expensive if there are lots of fields or `replace_make_shared` is called many times with the same parameters.
Reviewed By: ngorogiannis
Differential Revision: D19446260
fbshipit-source-id: c5c279ca5
Summary: This diff fixes the array access checking function for nested global arrays. We had assumed that RHS of `store` statement in SIL does not include array access expression, but that is not true: for global arrays, SIL can have statements like `*LHS = GlobalArray[n][m]`.
Reviewed By: ezgicicek
Differential Revision: D19300153
fbshipit-source-id: 256325642
Summary:
This diff gives semantics of `std::make_shared` as simple constructor, i.e., it changes function
call of `std::make_chared<C>(i)` to the constructor `C(i)`.
Reviewed By: ngorogiannis
Differential Revision: D19432338
fbshipit-source-id: 0d838e555
Summary: This adds a check for when developers use weakSelf (and maybe strongSelf) in a block, when there is no need for it, because it won't cause a retain cycle. In general there is an annotation in Objective-C for methods that take blocks, NS_NOESCAPE, that means that the passed block won't leave the scope, i.e. is "no escaping". So we report when weakSelf is used and the block has the annotation.
Reviewed By: ngorogiannis
Differential Revision: D19662468
fbshipit-source-id: f5ac695aa
Summary:
This attribute is given to parameters of methods that take Objective-C blocks to show that they will be used only in the current context and won't "escape" the context.
We translate it here, with the goal to use it in a new check later. The check is about not using weakSelf in non-escaping blocks, because retain cycles are not possible.
The translation is a bit complex because the annotation comes in the parameter of a method, but in the checker we will need it in the block. So we pass it around in the frontend from the translation of the method call to the translation context and on to the block expression and the block declaration afterwards.
Reviewed By: ngorogiannis
Differential Revision: D19600377
fbshipit-source-id: dd49539bd
Summary:
The difference from default set is mostly in:
1. turning on warnings 32..39 `unused X`,
2. turning off warning 66 `unused open!` since `open! IStd` is present
in pretty much all the files.
3. Non-exhaustive pattern match is now treated as a compilation
error.
Reviewed By: ngorogiannis, mityal
Differential Revision: D19646047
fbshipit-source-id: c84ba628a
Summary:
Refactor all occurences of `is_strict_mode` to use `NullsafeMode`
instead. This will allow introducing _local_ typechecking modes for
nullsafe in the follow up patches.
Reviewed By: ezgicicek
Differential Revision: D19639883
fbshipit-source-id: bdf535b66
Summary: Moving this big tuple to a record, because it's cleaner code, and I need to add another element in the next diff.
Reviewed By: ezgicicek
Differential Revision: D19640389
fbshipit-source-id: 86b1576a0
Summary:
Update from dune 1 to dune 2. The change was mostly straightforward
except that change to (modes ...) default to exe only was somewhat
unexpected.
Anyhow, with the community moving to dune 2 it's good to keep up.
Reviewed By: ngorogiannis
Differential Revision: D19605895
fbshipit-source-id: 1f9830de8
Summary: This will be eventually used in RacerD. Further diffs up the stack will migrate away the starvation-specific aspects.
Reviewed By: ezgicicek
Differential Revision: D19623392
fbshipit-source-id: a72650718
Summary:
Prevent returning a negative cost bound when calling `substring(begin_index, end_index)` when either is possible
- `begin_index < 0`
- `begin_index > end_index`
Instead, return unit cost since such cases either throw `IndexOutOfBoundsException ` at runtime or correspond to having two symbolic bounds that cannot be semantically compared.
Reviewed By: dulmarod
Differential Revision: D19619410
fbshipit-source-id: cf5e8cb7b
Summary:
The "access path" memory model (equal access paths iff equal object addresses) is suited to when aliasing occurs only at the roots (i.e. variables). When there is intentional aliasing in the middle of an access path, this model will miss the aliasing. For instance if `[x.f] == [y.g]`, then also `[x.f.h] == [y.g.h]`, but the latter access paths are unequal.
In Java, non-static inner classes consistently alias `this.this$0` inside an inner class, which points to the "parent" outer-class object. So if two inner-class objects (belonging to different inner classes) access `this(type:InnerClassA).this$0.f` and `this(type:InnerClassB).this$0.f` the equality will be missed (many other combinations exist). This isn't strictly due to the memory model -- any alias analysis would have to do some class invariant inference to detect this.
For this purpose `AccessPath.inner_class_normalize` exists (it replaces `this.this$0` with `this` of the appropriate type), but this breaks the invariant that we know which formal parameter is at the root (there may not even exist a `this` parameter if the method is static). So this was buggy.
Here we simply recursively remove the synthetic field prefix of the accesses list, while computing forwards the object type. This is only applied when we check aliasing across threads. This will also allow actuals/parameters substitutions (stacked diff) which normalisation was breaking.
Reviewed By: jberdine
Differential Revision: D19601455
fbshipit-source-id: 7e42667b6
Summary:
Rename DeclaredNonnull -> UncheckedNonnull, Nonnull -> StrictNonnull.
This is a preparatory step to introduce one more nullability option,
specifically CheckedNonnull to support local nullsafe mode.
Reviewed By: ngorogiannis, mityal
Differential Revision: D19619289
fbshipit-source-id: 12a3ff814
Summary:
This will help adoption of non-transitive strictification modes.
The main use-case we are aiming with this function is:
- During strictification, a method is made nullable, but it is not feasible to fix all
callers straight away. Suppressing nullsafe with an appropriate task or
comment will make it easier to unblock the strictification and move
forward.
Reviewed By: artempyanykh
Differential Revision: D19599098
fbshipit-source-id: 3769bbd3d
Summary:
1. I can not imagine a useful usecase for strict node, lets not
complicate the code.
2. Add docs.
Reviewed By: artempyanykh
Differential Revision: D19599091
fbshipit-source-id: 1d37a717f
Summary:
Reorder assertNotNull and assumeNotNull:
1. More useful methods go first.
2. More useful methods explicitly recommended over less useful.
Reviewed By: artempyanykh
Differential Revision: D19599084
fbshipit-source-id: 1b9ad9ef6
Summary: The translation of closures includes a load instruction for the captured variable, then we add that corresponding id to the closure. This doesn't correspond to an actual "use" of the captured variable in the source program though, and causes false positives. Here we remove the ids from the domain when that id is being added to a closure.
Reviewed By: ngorogiannis
Differential Revision: D19557352
fbshipit-source-id: 52b426011
Summary:
- Add `Nullsafe` annotation as a general mechanism to specify
type-checking behaviour for nullsafe.
- Document annotation params and provide usage examples with
explanations.
- Add tests to demonstrate the behaviour with different type-checking
modes.
No implementation is added. This diff serves as an RFC to hash out the
details before I dive into code.
Reviewed By: ngorogiannis
Differential Revision: D19578329
fbshipit-source-id: b1a9f6162
Summary: Keep the type name of the class as the key in the map constructed from class names to their methods in a file. This will be used later, and also why string?
Reviewed By: dulmarod
Differential Revision: D19557707
fbshipit-source-id: aa8569581
Summary: D19496263 had refactored cost to its own subdirectory. In this process, it introduced a perf degradation by bundling of collection and computation of constraints which relies on imperative union find. This diff fixes that by separating them.
Reviewed By: ngorogiannis
Differential Revision: D19598610
fbshipit-source-id: 0e466522d
Summary: Add the interface that will be used to lock Procedures.
Reviewed By: ngorogiannis
Differential Revision: D19580578
fbshipit-source-id: b5e334b18
Summary: `String.split(regexp)` returns an array that is split by the given regexp. If the regexp doesn't match, the original string is returned. Hence, the resulting array's length must be in `[1, max(1, n_u -1)]` where`n_u` is the upper bound of the string's length.
Reviewed By: dulmarod
Differential Revision: D19578318
fbshipit-source-id: 675af7376
Summary:
Adding a new check as part of the SelfInBlock checker, for cases when a strong pointer to self, that is not self, is captured in the block.
This will cover the following wrong scenarios:
1. strongSelf is a local variable in a block as it should be, and then it's used in a sub-block, in which case it's a captured variable.
2. weakSelf is defined without the weak attribute by mistake.
Reviewed By: ngorogiannis
Differential Revision: D19538036
fbshipit-source-id: 151871745
Summary:
In practice, condition redundant is extremely noisy and low-signal
warning (hence it is turned off by default).
This diff does minor tweaks, without the intention to change anything
substantially:
1/ Change severity to advice
2/ Change "is" to "might be"
3/ Describe the reason in case the origin comes from a method.
The short term motivation is to use 3/ for specific use-case: running nullsafe on codebase and
identify most suspicious functions (that are not annotated by often
compared with null).
Reviewed By: artempyanykh
Differential Revision: D19553571
fbshipit-source-id: 2b43ea0af
Summary:
This is/can be useful for future changes that make the reporting more
precise based on the exact origin.
See the follow up diff as an example.
Reviewed By: artempyanykh
Differential Revision: D19553567
fbshipit-source-id: c2b2c28a1
Summary: Currently, empty summaries are passed to the cluster callbacks. This is pointless and could potentially lead to recomputing already analysed summaries. This change passes only procnames to the callback, and `Ondemand` is used to load the summary or analyse as needed.
Reviewed By: jvillard
Differential Revision: D19544043
fbshipit-source-id: 28ab642c3
Summary: There is no need to provide type environments to cluster analysers, since the execution environment can be used to retrieve those on demand.
Reviewed By: jvillard
Differential Revision: D19543561
fbshipit-source-id: f9b064011
Summary: We were reporting strongSelf Not Checked when the variable was checked in a conditional, this diff fixes that.
Reviewed By: jvillard
Differential Revision: D19535943
fbshipit-source-id: f8e64e1b7
Summary: I noticed when looking into a false positive of strongSelf Not Checked, that there were some inconsistencies in the translation of if statements with an and, with an extra redundant join only if using a method in the condition that returned an object. So I could repro the problem and investigate and found the place of the inconsistency in the translation. This diff fixes it without changing things too much.
Reviewed By: jvillard
Differential Revision: D19518368
fbshipit-source-id: 47a6a778c
Summary:
The order by which the scheduler visits odd and even methods here
will determine if there is any report at all. This is a bad test
so remove.
Reviewed By: fgasperij
Differential Revision: D19535537
fbshipit-source-id: 6b64b0de9
Summary: Adding reporting to strongSelf Not Checked when strongSelf is passed to a method in a not explicitly nullable position.
Reviewed By: ezgicicek
Differential Revision: D19330872
fbshipit-source-id: 95871a70a
Summary: After receiving feedback about this, I'm changing the reporting of strongSelf Not Checked to only in cases where it can cause a crash. Here I'm adding reporting for field access, and removing general reporting. In a next diff, I'll also add reporting for passing strongSelf to methods in not explicitly nullable positions.
Reviewed By: skcho
Differential Revision: D19329842
fbshipit-source-id: 35beb2aa3
Summary:
The RestartScheduler needs to know if the worker finished it's task
because:
1. there was no more work to do or
2. found that a needed Procname was already taken (this part is not yet implemented)
This need was addressed by (i) making the functions that the workers execute
return a value of task_result.t intead of unit and (ii) adding a
constructor to the worker_message.t (FinishedTask).
Reviewed By: ngorogiannis
Differential Revision: D19467783
fbshipit-source-id: a76b02b6c
Summary:
1. One should use either a writer or a stream to send a response, but not both.
2. A response should be forwarded only if it was commited.
Both properties are extracted from API comments on classes in the servlet API.
Reviewed By: jvillard
Differential Revision: D19514568
fbshipit-source-id: 79f0257ed
Summary:
If data comes from an outer OutputStream, then this outer OutputStream
needs to be flushed before getting the byte array.
Reviewed By: jvillard
Differential Revision: D19514569
fbshipit-source-id: e3e025394
Summary: We were lacking this kind of test where one interface refines the nullability of the other.
Reviewed By: ngorogiannis
Differential Revision: D19514245
fbshipit-source-id: fa3e781f3
Summary: `cost.ml` is huge. Let's split it to its logical parts (basically creating new files for the modules that were already in `cost.ml`) and move all cost related files into `\cost` directory. While we are at it, let's add `mli` files too.
Reviewed By: jvillard
Differential Revision: D19496263
fbshipit-source-id: 45096db4c
Summary: Deadlocks often result in two reports if not deduplicated (two traces), so there is some logic for doing that. Locks recently became an opaque type, with the `get_access_path` loophole supporting that deduplication ordering. Fix that here and remove `Lock.get_access_path`.
Reviewed By: jvillard
Differential Revision: D19465223
fbshipit-source-id: b597e3c65
Summary:
This is a common enough case to make error message specific.
Also let's ensure it's modelled.
Reviewed By: artempyanykh
Differential Revision: D19431899
fbshipit-source-id: f34459cb3
Summary:
The previous diff changes the message for params case, this one handles
return.
Reviewed By: artempyanykh
Differential Revision: D19430706
fbshipit-source-id: f897f0e56
Summary:
This diff gets global constant array values from their initializers. The `find_global_array` function is
added to memory domain, which finds values of global array locations during the ondemand value
generation.
Reviewed By: ngorogiannis
Differential Revision: D19300143
fbshipit-source-id: 7b0b84c42
Summary: Use more informative method names, and add comments explaining the logic behind each test. Correct two cases which are FPs instead of legitimate reports.
Reviewed By: artempyanykh
Differential Revision: D19465227
fbshipit-source-id: 29332e2b9
Summary: Change `MayBlock` and `StrictModeCall` constructors from taking a string to a `Procname.t`, which was the sole source of that string anyway.
Reviewed By: ezgicicek
Differential Revision: D19465226
fbshipit-source-id: e3ed6ef88
Summary:
If a race exists in two or more overloads of the same method and we use only the class and method name in the report text, then the current bug hashing algorithm will identify the two reports as duplicates.
To avoid this, the report had the class, method and list of type parameters. This is unreadable, however, and redundant (the report is already located within the method in question). So at the risk of duplicates, use only class+method names.
Also, fix a bug in `Procname.pp_simplified ~withclass` where `withclass` was ignored for C++/ObjC methods.
Now:
> Read/Write race. Non-private method `FrescoVitoImageSpec.onCreateInitialState(...)` indirectly reads with synchronization from `factory.AnimatedFactoryProvider.sImpl`. Potentially races with unsynchronized write in method `FrescoVitoImageSpec.onEnteredWorkingRange(...)`.@ [Litho components are required to be thread safe because of multi-threaded layout](https://fburl.com/background-layout). Reporting because current class is annotated `MountSpec`, so we assume that this method can run in parallel with other non-private methods in the class (including itself).
Before
> Read/Write race. Non-private method `void FrescoVitoImageSpec.onCreateInitialState(ComponentContext,StateValue,StateValue,Uri,MultiUri,ImageOptions,FrescoContext,Object,ImageListener)` indirectly reads with synchronization from `factory.AnimatedFactoryProvider.sImpl`. Potentially races with unsynchronized write in method `FrescoVitoImageSpec.onEnteredWorkingRange(...)`.@ [Litho components are required to be thread safe because of multi-threaded layout](https://fburl.com/background-layout). Reporting because current class is annotated `MountSpec`, so we assume that this method can run in parallel with other non-private methods in the class (including itself).
Reviewed By: artempyanykh
Differential Revision: D19462277
fbshipit-source-id: aebc20d89
Summary:
Currently, impurity analysis is oblivious to skipped functions which might e.g. return a non-deterministic value, write to memory or have some other side-effect. This diff fixes that by relying on Pulse's skipped functions to determine impurity. Any unknown function which is not modeled to be pure is assumed to be impure.
This is a heuristic. We could have assumed them to be pure by default as well.
Reviewed By: jvillard
Differential Revision: D19428514
fbshipit-source-id: 82efe04f9
Summary:
Let's collect the list of all skipped functions with a `proc_name` but no summary in Pulse's memory. This will be useful for the impurity analysis later (next diff).
Concretely, we extend Pulse's domain with a map from skipped calls to their respective traces. For efficiency, we only keep a single trace per skipped call.
For impurity analysis, tracking skipped calls in Pulse allows us to rely on Pulse's strong memory model to get rid of infeasible paths as opposed to creating an independent checker which wouldn't be able to do that.
Reviewed By: jvillard
Differential Revision: D19428426
fbshipit-source-id: 3c5e482c5
Summary:
As suggested by Ilya, the current message can be improved in a way that
it can contain more clear action. I also added artempyanykh's explanation at the
end of message to provide an additional justification from common sense
perspective.
But most importantly, the previous message was missing a space which is
eye bleeding, how come haven't I noticed this before, I can't stand it
OMG.
Reviewed By: artempyanykh
Differential Revision: D19430271
fbshipit-source-id: dd31f7adb
Summary:
Inferbo analyzed some program points unreachable incorrectly, because of unsound semantics of band
operator, which did not handle the case when given parameters are pointer values.
Reviewed By: ngorogiannis
Differential Revision: D19392705
fbshipit-source-id: dd590508c
Summary:
This diff revises the generation of unknown value. If the type of the unknown value generating is
int, it does not add the "Unknown" pointer/array value.
Reviewed By: ngorogiannis
Differential Revision: D19392696
fbshipit-source-id: e1b3c9a3a
Summary: In impurity analysis, pick up the pulse summary rather than re-analyzing. Re-order the checkers so that we first analyze pulse.
Reviewed By: jvillard
Differential Revision: D19448296
fbshipit-source-id: 2987fa848
Summary:
This diffs does: (1) move `get_formals` to `BufferOverrunUtils` (2) use separate `get_formals` in
`BufferOverrunChecker`, in order to simplify the following diff.
Reviewed By: jvillard
Differential Revision: D19432280
fbshipit-source-id: bfb4df118
Summary:
The restart scheduler will now have two phases:
1. Analyze all the procs obtained from the sources.
2. Run the FileScheduler on the sources.
The second step aims to analyze only the File level analyzers
requirements.
Reviewed By: ngorogiannis
Differential Revision: D19430244
fbshipit-source-id: b4f9ee69b
Summary: `dune build check` will compile all the ml files much quicker than `dune build`
Reviewed By: jvillard
Differential Revision: D19427864
fbshipit-source-id: 5221d32bc
Summary:
Introduce a new notion of equality for comparing abstract addresses in distinct threads:
```
(** Abstract address for a lock. There are two notions of equality:
- Equality for comparing two addresses within the same thread/process/trace. Under this,
identical globals and identical class objects compare equal. Locks represented by access paths
rooted at method parameters must have equal access paths to compare equal. Paths rooted at
locals are ignored.
- Equality for comparing two addresses in two distinct threads/traces. Globals and class objects
are compared in the same way, but locks represented by access paths rooted at parameters need
only have equal access lists (ie [x.f.g == y.f.g]). This allows demonically aliasing
parameters in *distinct* threads. This relation is used in [may_deadlock]. *)
```
Reviewed By: skcho
Differential Revision: D19347307
fbshipit-source-id: 9f338731b
Summary:
This diff avoids that `array_sizeof` returns bottom value when given Java enum values, which
introduced unreachable code inadvertently.
Reviewed By: ngorogiannis
Differential Revision: D19409077
fbshipit-source-id: 2816fd995
Summary: Access expressions can appear in casts, or sometimes other constructors, inside a `HilExp.t`. Extraction of the access expression can ignore those wrappers. Introduce a single function for doing that throughout the analyser.
Reviewed By: ezgicicek
Differential Revision: D19410673
fbshipit-source-id: a724cb466
Summary:
The property SkipAfterRemove already had a test, but not for
intra-procedural violations. This adds a test for that case.
Reviewed By: ngorogiannis
Differential Revision: D19330471
fbshipit-source-id: 1dd1c3ad7
Summary:
Now we can either disable it, or enable it only.
For integrations that disable it, this will allow to enable it for
NullsafeStrict classes without enabled it fully
Reviewed By: artempyanykh
Differential Revision: D19409131
fbshipit-source-id: a2b1fe650
Summary: This diff implements this for Field Not Initialized check
Reviewed By: artempyanykh
Differential Revision: D19393989
fbshipit-source-id: cf60e8d53
Summary: add subdirectories so that we can run each java file against its own topl properties
Reviewed By: rgrig
Differential Revision: D19347302
fbshipit-source-id: 562830774
Summary:
This diff does it for nullable dereference and assignment violations
rules which happen under NullsafeStrict case.
Follow up are to make the same for inheritance and field initializer
violations.
Possible follow up includes making error message more specific and
articulare this this is a nullsafe strict mode.
Reviewed By: artempyanykh
Differential Revision: D19392916
fbshipit-source-id: 2554ac7a7
Summary: In whole-program mode, analysing a method requires analysing first all constructors of the same class. This is not needed in normal mode, so gate that computation under `starvation_whole_program` for efficiency.
Reviewed By: artempyanykh
Differential Revision: D19393412
fbshipit-source-id: 2277e6b5e
Summary:
The RestartScheduler now uses it's own of_list function to build a
ProcessPool.TaskGenerator to be able to use a queue as the underlying
data structure.
Reviewed By: ngorogiannis
Differential Revision: D19390055
fbshipit-source-id: d57b493b7
Summary:
Previously, _override resolution_ considered only the number of
arguments. This led to many FPs in nullsafe's _Inconsistent Subclass
Annotation_ check.
Current version also checks that argument types match. However, we
still don't handle type parameters and erasure, so in this sense the
rules are incomplete.
Reviewed By: ngorogiannis, mityal
Differential Revision: D19393201
fbshipit-source-id: a0c75b8dd
Summary:
That was used only by the now-defunct Java Buck integration without
genrules.
Reviewed By: ngorogiannis
Differential Revision: D19176545
fbshipit-source-id: 8253b614a
Summary:
You should use `--buck-java` instead, which uses the new "genrule
master" integration.
This diff makes it impossible to select the previous integration.
Upcoming diffs will clean up the resulting dead code.
This also make infer fail hard when no buck mode is specified in a buck
capture command, eg `infer -- buck build //foo:foo`. The reason is that
we need to choose between 3 incompatible integrations and making any of
them the default will confuse at least one person in the future.
Reviewed By: ngorogiannis
Differential Revision: D19176391
fbshipit-source-id: 707d18b50
Summary:
These changes introduce the RestartScheduler which for now is a copy of the FileScheduler:
- added it as a possible argument to the recently added `--scheduler` option.
- made the necessary changes in `InferAnalyze` to call it if it was chosen.
Reviewed By: ngorogiannis
Differential Revision: D19373505
fbshipit-source-id: 98f065057
Summary:
`diff` by default doesn't colorize its output, while `git diff`
does. Since we already depend on git it seems like a safe change and
brings some quality of life improvement while working with tests.
Reviewed By: ngorogiannis
Differential Revision: D19374691
fbshipit-source-id: 8872b527c
Summary:
`infer/tests/infer.make` unconditionally runs
```
$(call check_no_duplicates,$(INFER_OUT)/duplicates.txt)
```
as a part of `test` target. This requires the flag to be set to
produce `duplicates.txt`.
`clang.make` does this, but `java.make` was missing the flag which led
to an error on running `make tests` within
e.g. `tests/codetoanalyze/java/nullsafe-default`:
```
grep: infer-out/duplicates.txt: No such file or directory
```
Reviewed By: mityal
Differential Revision: D19373857
fbshipit-source-id: c2fcbe9dd
Summary:
This diff avoids that null-retuned path's abstract value ruins that of non-null-returned path.
What this diff does is: when joining two abstract states, one is null-return-path and the other is
non-null-return-path (`return obj;`), it keeps the method calls of `obj` from the
non-null-return-path.
While this design is unsound, I think it should work in practice.
Reviewed By: ezgicicek
Differential Revision: D19348313
fbshipit-source-id: cf5d0f3ff
Summary: This diff captures global initializers ondemand, like we do for functions defined in headers.
Reviewed By: ezgicicek
Differential Revision: D19346947
fbshipit-source-id: 05174e6a4
Summary: The type-name definition for Java can be potentially improved (eg increase sharing, or comparison speed, much like `QualifiedCppName`) by switching away from `Mangled.t` which is essentially a string. First step is to abstract the type.
Reviewed By: jberdine
Differential Revision: D19087508
fbshipit-source-id: 91a81f63b
Summary: Now that we have the kind of lock stored (global/class obj/path rooted at parameter), use it for comparison/equality, while ignoring the root variable of the access path, which is only used for printing.
Reviewed By: skcho
Differential Revision: D19346801
fbshipit-source-id: c65661dc6
Summary: Add command line option that directs racerd to treat all return values from unknown code (including abstract methods) as owned objects. This is essentially treating return values with full angelicism
Reviewed By: artempyanykh
Differential Revision: D19368375
fbshipit-source-id: 6a10153fa
Summary:
Java treats switch on nullables in a non-obvious way (throws an NPE
surprise) so lets have a decidated test exactly for this.
Reviewed By: ngorogiannis
Differential Revision: D19371280
fbshipit-source-id: d9867b6d6
Summary:
As part of enabling substitution of arguments into lock names in summaries, and potentially changing the equality relation over locks (so that we can eventually make equal `x.f.g` and `this.f.g` for example), the lock type needs to be elaborated so that the root of an access path is classified into:
- global variables are clearly separated from all other classes (invariant to substitution)
- class objects are also separate (also invariant, but identified by type)
- all other parameters remember their positional index (so that substitution becomes easier)
For now the comparison/equality is kept identical, so as to make easy CI comparisons.
Reviewed By: skcho
Differential Revision: D19232577
fbshipit-source-id: 0cd6a43db
Summary: Before going to a new lock representation that will allow, eg, substitutions with arguments for parameters on method calls, make the interface of the lock type abstract to ease the transition.
Reviewed By: mityal
Differential Revision: D19309359
fbshipit-source-id: 5277357ee
Summary:
Demonstrate that the per-file type environments don't prevent
the deadlock report here. The fear was that when the analyser
tries to locate the methods of the endpoint class, it might fail to
do so because the types might be stored in different type
environments (per file).
Reviewed By: mityal
Differential Revision: D19225908
fbshipit-source-id: 097e4aeea
Summary: This diff use actuall call path in the cost results instead of `class name + method name`.
Reviewed By: ngorogiannis
Differential Revision: D19194969
fbshipit-source-id: b72018586
Summary:
Model array length in Java as returning an unknown interval [0, +inf] for now.
Ideally, we can deal with the size in a more precise manner in the future like in InferBo.
Reviewed By: skcho
Differential Revision: D19312123
fbshipit-source-id: 8c51059a4
Summary:
This diff does a refactoring on the function that gets summary from DB,
`get_proc_summary_and_formals`.
* It separates the function into `get_summary` and `get_formals`, and
* renames `Checker.get_proc_summary` to `get_checks_summary`
Reviewed By: ngorogiannis
Differential Revision: D19300136
fbshipit-source-id: d28eaf16d
Summary:
This diff creates missing result directories in its running.
The problem was that `infer-out/captured` and its sub-directories were not ready at the step 3 below, which crashed with exceptions.
1. `infer capture -- [target build]`
2. `infer analyze --merge`
3. `infer analyze --merge --debug --reanalyze --procedures-filter '.*foo.*'`
Reviewed By: ngorogiannis
Differential Revision: D19274672
fbshipit-source-id: af84000d7
Summary: Pulse doesn't care about exceptions yet. With Exceptional CFG, java analysis takes a lot of time due to having many disjuncts. Let's use Normal CFG for now.
Reviewed By: jvillard
Differential Revision: D19194479
fbshipit-source-id: f94bb6078
Summary:
Kill these tests as that mode is not supported anymore.
Note that :
- "buck_cross_module" was effectively already dead since it's not hooked up to `make test`
- "buck" is actually another copy of "genrule" so wasn't testing anything
Reviewed By: ngorogiannis
Differential Revision: D19176166
fbshipit-source-id: b5bb90448
Summary:
This changes how we select amongst our (currently) 4 Buck integrations
for Java and clang, as well as how the user's choice is reflected by the
Config module.
The old command line interface is still supported but is now deprecated.
The changes in how to select each integration are:
- clang via "flavors", activated with `--flavors`, now with `--buck-clang`
- clang via "compilation DB", activated with `--buck-compilation-database`, unchanged
- Java via "genrule", activated with `--genrule-master-mode`, now with `--buck-java`
- Java "without genrules", used to be activated by *not specifying any other Buck mode*, unchanged
Instead of various `Config` flags corresponding to the previous CLI that
are allowed in any combination of `flavors`,
`buck_compilation_database`, `genrule_master_mode`, `Config` now exposes
a single `buck_mode` datatype. This allows, eg, `flavors` to override
`buck_compilation_database` if needed. It will also make it easier to
get rid of the old "Java without genrules" integration in a later diff
(see inline comments).
Reviewed By: ngorogiannis
Differential Revision: D19175686
fbshipit-source-id: 29b3831be
Summary:
We cannot necessarily know if the previous capture phase was java or
clang, hence to be on the safe side we should always merge the global
tenvs when merging Buck targets.
Reviewed By: ngorogiannis
Differential Revision: D19175685
fbshipit-source-id: 8e7492e14
Summary:
This generates a lot of code (the datatype is not trivial) and
transitively requires other datatypes to have compare functions, just so
we can then ask whether the mode is "Analyze".
Reviewed By: ngorogiannis
Differential Revision: D19164589
fbshipit-source-id: be6a12b41
Summary:
Infer would crash when started in a context where it already had
children it didn't know about.
Reviewed By: martintrojer
Differential Revision: D19177589
fbshipit-source-id: 0c8831597
Summary:
In order to improve the impurity analysis, this diff adds models for
- `hasNext()` and - `Object.equals()` modeled as returning a non-deterministic value (havoc_id)
- `next()` modeled as `StdVector.get` with a fresh index
- `iterator` modeled as just returning the underlying list
Reviewed By: jvillard
Differential Revision: D19177392
fbshipit-source-id: 0babb037a
Summary:
This diff updates the relation between iterator (offset) and integer value not only at
assignments (`x += 1`), but also at function calls (`foo()`) that increase integer values by one in
their side effects.
Reviewed By: ezgicicek
Differential Revision: D19163214
fbshipit-source-id: 47e52f939
Summary: This diff extends the domain to express the relation between iterator's offset and integer value.
Reviewed By: ezgicicek
Differential Revision: D19143670
fbshipit-source-id: 6223bc934
Summary:
Old versions of sawja/javalib got the line numbers slightly wrong. The workaround was to do a regexp search in the source file for the right line.
My understanding is that this is no longer necessary. This diff removes it.
Reviewed By: jvillard
Differential Revision: D19033415
fbshipit-source-id: 2da19d66d
Summary:
From a set of classes [classes], JClasspath.load_program builds a
program. This program contains (implicitely) the set of active classes
and we should not keep using the set [classes] and the generated
program as the same time, because they may not be synched
anymore. (In particular, in a forthcoming diff load_program may add
new classes during program construction)
This patch forces this discipline. It adds a [in_classes]
query for programs and makes sure we use it everywhere we should.
Reviewed By: ngorogiannis
Differential Revision: D18833335
fbshipit-source-id: a522f320c
Summary:
This is an optimization. We ask the user to tell us which states are nondeterministic, and we
generate code that handle nondeterminism only for those states. It is common for only one state per
TOPL property to be nondeterministic. This speeds up the biabduction-analysis of the monitor by a
factor of ~10. But, using the monitor is only a little faster.
Facebook
Reviewed By: jvillard
Differential Revision: D19160286
fbshipit-source-id: 4dd39769a
Summary:
The re-execution phase uses only the preconditions found by the footprint phase. Thus, Infer should
not spend time optimizing POSTconditions. These are used only to find custom errors.
Reviewed By: jvillard
Differential Revision: D19035495
fbshipit-source-id: d52b82733
Summary:
Move most of IR/Sil.ml into a new file biabduction/Predicates.ml to
reflect the fact that they are only useful for the biabduction analysis.
Unfortunately this is a huge change.
I tried to keep the change to a minimum, it's mostly about doing
s/Sil/Predicates/ in lots of places but sometimes I used the trick of
specifying parameters or return value types to avoid specifying the
module altogether. This isn't done consistently because there were just
too many places to change for poor me.
Reviewed By: ngorogiannis
Differential Revision: D19158530
fbshipit-source-id: d6dbcfe72
Summary:
hmmhmmhmm
Also `Sil.hpred` is going to move out of IR/ in a few diffs, into
biabduction/ where it belongs.
Reviewed By: ngorogiannis
Differential Revision: D19158535
fbshipit-source-id: e2a889ee2
Summary:
Part of making Sil.ml about SIL only.
In order to not introduce a dependency istd/Pp -> base/Config, the
utilities in Pp don't know when to introduce "diff" colours. Fix it by
wrapping them in Sil using the Config option. (we may want to just kill
that option at some point).
Similarly, move stuff from Io_infer to Pp.
Reviewed By: ngorogiannis
Differential Revision: D19158534
fbshipit-source-id: 8110cb7f9
Summary:
In the previous code, it removed non-build-called method calls. For example, it was like
```
{non-build-called: {prop1, prop3}}
{build-called: {}}
b.build();
{non-build-called: {}}
{build-called: {prop1, prop3}}
```
However, this behavior introduced a false positive when there is multiple builders that point to the
same abstract object and `build` is called one by one.
This diff changes the semantics to keep the method calls of non-build-called at `build` calls.
Reviewed By: ezgicicek
Differential Revision: D19144525
fbshipit-source-id: e2ace127f
Summary:
At some point, code reformatting moved the comment
in a place where it didn't make sense.
Reviewed By: ngorogiannis
Differential Revision: D19035489
fbshipit-source-id: 146ad8a15
Summary:
This applies some simplifications that were previously
done after footprint (and therefore lost), and some
simplifications that require looking at both pre and
post.
Reviewed By: ngorogiannis
Differential Revision: D19035494
fbshipit-source-id: bad79534a
Summary:
This havocs event data, so that biabduction doesn't try to
track what was the last event processed by the monitor
(which is redundant as long as the state of the monitor
is tracked).
Reviewed By: ngorogiannis
Differential Revision: D19035491
fbshipit-source-id: a1c75daae
Summary:
Don't instrument SIL when we can determine statically that
biabduction symexec would be a no-op.
Reviewed By: ngorogiannis
Differential Revision: D19116849
fbshipit-source-id: 4d25462a3
Summary: It is confusing to report missing props at the the beginning of the method (especially when there are many components created or when the method has many lines). Let's report them at create methods to better contextualize. Also, make the missing prop bold.
Reviewed By: skcho
Differential Revision: D19141135
fbshipit-source-id: a23d2e7c9
Summary:
It used to recompute the number of arguments during
instrumentation, which is unnecessary.
Reviewed By: ngorogiannis
Differential Revision: D19035492
fbshipit-source-id: 0f58c8ae7
Summary:
In addition to
state1 -> state2: pattern
one can now also write
state1 -> state2: pattern if condition
where "condition" is some conjunction of comparisons (==,<,>) that
involve variables bound by "pattern", registers of the automaton, and
constants.
Reviewed By: ngorogiannis
Differential Revision: D19035496
fbshipit-source-id: 6f6e6a9be
Summary:
The one with Object as the second param is in `com.facebook.common.internal`.
The one with String as the second param is in `com.facebook.common.preconditions`.
(Don't ask :) )
Reviewed By: artempyanykh
Differential Revision: D19087334
fbshipit-source-id: ba02c9101
Summary:
According to Java semantics, they are always non-null.
Internally they are represented as static fields, so they have
DeclaredNonnull nullability, which means NullsafeStrict mode would
refuse to use them without strictification.
Lets teach nullsafe that these guys are non-nullables.
See also FN in test case.
Reviewed By: ngorogiannis
Differential Revision: D19024547
fbshipit-source-id: 8c120fa50
Summary:
Inferbo does not use the external relational domains, apron and elina. At some point, the parts of
inferbo using them were broken and they do not seem to be fixed easily in the near future. Let's
remove them and keep the code base cleaner.
Reviewed By: jvillard
Differential Revision: D19022905
fbshipit-source-id: e0eafe79f
Summary:
We do not have the create method in the trace which makes it difficult to understand
- inter-procedural issues where create and prop setting are in different methods
- there are multiple create-build chains in a method
Let's add the create to the beginning of the trace. Moreover, let's simplify the prop printing to make traces easier to understand.
Reviewed By: ngorogiannis
Differential Revision: D19020213
fbshipit-source-id: 7f8a5d4ec
Summary: The new domain is much better than the old one. Let's kill the old one (along with old litho tests) and simplify things.
Reviewed By: skcho
Differential Revision: D18959627
fbshipit-source-id: df77ae20e
Summary:
- Remove `to_flat_string` as there is `get_field_name` that unambiguously does the same thing.
- Make `pp` print only the field in all languages.
- Fix `to_full_string` so that it has unified behaviour across java/clang and so that it doesn't print `class Foo.x`, but rather `Foo.x`.
Reviewed By: ezgicicek
Differential Revision: D18963033
fbshipit-source-id: e2c803c7d
Summary:
Remove Clang and Java submodules of Typ.Fieldname. They are unnecessary and they reflect a fake dichotomy: there is only one fieldname type. To distinguish between fields of Java classes and other C constructs, there is a helper function provided, but the idea is simple: obtain the class type the field belongs to, and check if it's a Java class.
This diff still preserves behaviour, but removes as many functions as possible from the interface, to leave a small surface.
Reviewed By: mityal
Differential Revision: D18962423
fbshipit-source-id: ffe6933ee
Summary: Unify treatment of Java and Clang fieldnames. Now a field is a struct with a class type-name and a string-field name. This diff is still behaviour preserving.
Reviewed By: jvillard
Differential Revision: D18953549
fbshipit-source-id: 8cae0d104
Summary: This function allows any string, and in particular empty class names. As a first step eliminate it in favour of a function that forces the caller to specify distinct class and field names. It turns out that the frontend already has them, so it saves effort along the way.
Reviewed By: jvillard
Differential Revision: D18953136
fbshipit-source-id: ff3cdfda5
Summary:
In order to handle the example added:
changed domain of `MethodCalled`
from `CreatedLocation -> (IsBuildCalled X IsChecked X Set(MethodCall))`
to `(CreatedLocation X IsBuildCalled) -> (IsChecked X Set(MethodCall))`
This avoids joining of two method calls where one is build-called and the other is not, e.g.,
```
if(b) {
o.build();
} else {
// no build call
}
```
changed domain of `NewDomain`
from `Created X MethodCalled`
to `(Created X MethodCalled) X (Created X MethodCalled)`
One is for no returned memory and the other is returned memory. This keeps precision some join
points of branches, e.g.,
```
if(b) {
return;
} else {
// no return
}
```
Reviewed By: ezgicicek
Differential Revision: D18909768
fbshipit-source-id: c39d1a1ef
Summary:
A plus is a plus, no need to give up when +/- is about pointers. This
gets rid of some false positives involving pointer arithmetic.
However, the problem remains if we make things a bit more
inter-procedural. This is documented in an added test.
Reviewed By: ezgicicek
Differential Revision: D18932877
fbshipit-source-id: 4ad1cfe72
Summary: It is not used anywhere and there are no plans to revive it. Kill it!
Reviewed By: skcho
Differential Revision: D18934719
fbshipit-source-id: b9b069b96
Summary: Under the buck/java integration, the classpath is propagated to all infer command lines. Logging that leads to huge waste and full disks. Make the logging conditional to debug mode.
Reviewed By: skcho
Differential Revision: D18934088
fbshipit-source-id: 7e2f410f5
Summary:
The `Typ.FIeldname` module has many issues. Among those:
- It has 5 different string/printing functions and most of them do radically different things in Java and in Clang.
- There is no type safety: creating a Clang field and calling a Java function on it will lead to a crash (`rindex_exn` etc, there are usually no dots in Clang fields).
- It uses a single string for Java fields, containing the package, the class and the field, e.g., `java.lang.Object.field`. This is wasteful, because
- there is no sharing of strings for packages/classes, and,
- string operations need to be performed every time we need the field or the class or the package alone.
This diff preserves the behaviour of the module's interface, so the API problems remain.
However, by using a saner representation for Java fields we can get small performance and large memory gains (the type environment in Java is much smaller, about 30-40%).
In addition, many functions on clang fields would previously do string manipulations (look for `.` and split on it) before returning the final field unchanged -- now they use the type of the field for that.
Reviewed By: jvillard
Differential Revision: D18908864
fbshipit-source-id: a72d847cc
Summary:
That class does some complicated accounting of memory that depends on
whether the string is "small", "medium" or "large". In the latter case
it does its own ref-counting and copy-on-write to save memory, and that
trips up pulse. Pretending all strings are small avoids that issue.
Reviewed By: skcho
Differential Revision: D18909030
fbshipit-source-id: 1c14d909b
Summary:
Including the current call state is useful because the contradiction
sometimes refers to abstract values that have been materialised since
the last call state so we cannot make sense of them unless we print the
current call state.
Reviewed By: skcho
Differential Revision: D18908424
fbshipit-source-id: 297f397a6
Jules Villard
Dulma Churchill
Sungkeun Cho
Artem Pianykh
Nikos Gorogiannis
Ezgi Çiçek
Radu Grigore
David Pichardie
Fernando Gasperi Jabalera
Mitya Lyubarskiy
Martin Trojer
Kfir Schindelhaim
Artem Pyanykh