Shared: Re-factor summary, source and sink model generators into separate modules.#19382
Conversation
github-actions
bot
added
C#
C++
documentation
Java
Rust
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| private import semmle.code.cpp.dataflow.internal.FlowSummaryImpl as FlowSummaryImpl | ||
| private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific | ||
| private import semmle.code.cpp.dataflow.new.TaintTracking | ||
| private import semmle.code.cpp.dataflow.new.TaintTracking as Tt |
Check warning
Code scanning / CodeQL
Names only differing by case Warning
| private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific | ||
| private import semmle.code.cpp.dataflow.new.TaintTracking | ||
| private import semmle.code.cpp.dataflow.new.TaintTracking as Tt | ||
| private import semmle.code.cpp.dataflow.new.DataFlow as Df |
Check warning
Code scanning / CodeQL
Names only differing by case Warning
michaelnebel
force-pushed
the
shared/modelgenrefactor
branch
2 times, most recently
from
4fea2b0 to
fbe0394
Compare
michaelnebel
added
the
no-change-note-required
…rate param modules.
michaelnebel
force-pushed
the
shared/modelgenrefactor
branch
from
fbe0394 to
c9d01bc
Compare
MathiasVP
left a comment
MathiasVP
left a comment
There was a problem hiding this comment.
The C/C++ parts LGTM 👍
paldepind
left a comment
paldepind
left a comment
There was a problem hiding this comment.
The Rust changes LGTM :)
|
DCA for Java looks fine (workflow errors for C#). I think that is acceptable for confidence in the changes. |
|
The DCA retry for C# appears to have succeeded (modulo some other workflow failure that I don't understand). It appears that the re-factor doesn't impact performance. |
|
Merging now before we get more merge conflicts. |
|
DCA for Java looks fine (workflow errors for C#). I think that is acceptable for confidence in the changes. |
|
The DCA retry for C# appears to have succeeded (modulo some other workflow failure that I don't understand). It appears that the re-factor doesn't impact performance. |
|
Merging now before we get more merge conflicts. |
3 participants
In this PR we re-write the model generator into separate (nested) parameterized modules. Prior to this change it was only possible to create a model generator that could create summaries, neutrals, source and sinks. If only summary/neutral models are of interest, one still needed to provide all the source and sink model generation related parameters (as these were included in the input signature for model generator). With this change the model generator is turned into a series of nested parameterized modules.
That is, to create a model generator one now needs to
Furthermore, in this PR we sprinkle sensible predicate defaults for some of the model generator input predicates.