division-by-zero on float #8233
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
kroening
requested review from
tautschnig,
peterschrammel and
remi-delmas-3000
as code owners
kroening
force-pushed
the
float-div-by-zero
branch
from
9335e6b to
87ffbc8
Compare
kroening
force-pushed
the
float-div-by-zero
branch
4 times, most recently
from
b4f949a to
2f2e163
Compare
kroening
force-pushed
the
float-div-by-zero
branch
from
2f2e163 to
b733273
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## develop #8233 +/- ##
===========================================
- Coverage 79.64% 79.64% -0.01%
===========================================
Files 1684 1684
Lines 195670 195684 +14
===========================================
+ Hits 155841 155852 +11
- Misses 39829 39832 +3 ☔ View full report in Codecov by Sentry. |
kroening
force-pushed
the
float-div-by-zero
branch
2 times, most recently
from
e3f0ba9 to
16ea284
Compare
tautschnig
reviewed
Mar 10, 2024
| @@ -15,7 +15,7 @@ int main() | |||
| #pragma CPROVER check push | |||
| #pragma CPROVER check disable "bounds" | |||
| #pragma CPROVER check disable "pointer" | |||
| #pragma CPROVER check disable "div-by-zero" | |||
There was a problem hiding this comment.
This regression test is called "...enable_all," so shouldn't the new check be enabled in addition rather than replace an existing one?
There was a problem hiding this comment.
There isn't any integer division, so that wouldn't have any effect.
There was a problem hiding this comment.
Ah, but then this test should be extended to have one (the test is "enable_all", so we should be testing all of them).
tautschnig
approved these changes
Mar 22, 2024
| @@ -15,7 +15,7 @@ int main() | |||
| #pragma CPROVER check push | |||
| #pragma CPROVER check disable "bounds" | |||
| #pragma CPROVER check disable "pointer" | |||
| #pragma CPROVER check disable "div-by-zero" | |||
There was a problem hiding this comment.
Ah, but then this test should be extended to have one (the test is "enable_all", so we should be testing all of them).
kroening
force-pushed
the
float-div-by-zero
branch
from
16ea284 to
ffaf0ee
Compare
|
Test extended by adding an integer division by zero. |
This separates the division-by-zero check for floating-point operations from the division-by-zero check for integers. The new division-by-zero check for floating-point operations is off by default, and is enabled by --float-div-by-zero-check. The case for doing so is weak. ISO/IEC 9899:2021 (C21) and predecessors clealy state (Sec 6.5.5 par 5) "In both operations, if the value of the second operand is zero, the behavior is undefined." However, Annex F (IEC 60559 floating-point arithmetic) states that implementations that define __STDC_IEC_559__ must implement IEC 60559 division, which clearly defines the behavior when dividing floating-point numbers by zero. This behavior can be observed on all architectures that we support.
kroening
force-pushed
the
float-div-by-zero
branch
from
ffaf0ee to
e00a1bd
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This separates the division-by-zero check for floating-point operations from the division-by-zero check for integers. The new division-by-zero check for floating-point operations is off by default, and is enabled by
--float-div-by-zero-check.The case for doing so is weak. ISO/IEC 9899:2021 (C21) and predecessors clealy state (Sec 6.5.5 par 5) "In both operations, if the value of the second operand is zero, the behavior is undefined."
However, Annex F (IEC 60559 floating-point arithmetic) states that implementations that define
__STDC_IEC_559__must implement IEC 60559 division, which clearly defines the behavior when dividing floating-point numbers by zero. This behavior can be observed on all architectures that we support.