Fix segfault in removeVertexNode due to negative hash
#94
Conversation
This file contains hidden or 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
…gned The hash in `_hashVertex` produces out of range values if `(vertex->lat + vertex->lon)` was negative, because it would produce a negative module (such as -1) and then cast it to an unsigned integer type. This index would then cause a segfault. testH3SetToLinkedGeo.c is also updated to not `calloc` a fixed size of 1.
jogly
approved these changes
Jul 18, 2018
| @@ -29,130 +29,123 @@ H3Index* makeSet(char** hexes, int numHexes) { | |||
| BEGIN_TESTS(h3SetToLinkedGeo); | |||
|
|
|||
| TEST(empty) { | |||
| LinkedGeoPolygon* polygon = calloc(1, sizeof(LinkedGeoPolygon)); | |||
nrabinowitz
approved these changes
Jul 18, 2018
isaachier
approved these changes
Jul 18, 2018
| @@ -29,130 +29,123 @@ H3Index* makeSet(char** hexes, int numHexes) { | |||
| BEGIN_TESTS(h3SetToLinkedGeo); | |||
|
|
|||
| TEST(empty) { | |||
| LinkedGeoPolygon* polygon = calloc(1, sizeof(LinkedGeoPolygon)); | |||
| LinkedGeoPolygon polygon; | |||
| @@ -70,7 +70,7 @@ void destroyVertexGraph(VertexGraph* graph) { | |||
| uint32_t _hashVertex(const GeoCoord* vertex, int res, int numBuckets) { | |||
| // Simple hash: Take the sum of the lat and lon with a precision level | |||
| // determined by the resolution, converted to int, modulo bucket count. | |||
| return (uint32_t)fmod((vertex->lat + vertex->lon) * pow(10, 15 - res), | |||
| return (uint32_t)fmod(fabs((vertex->lat + vertex->lon) * pow(10, 15 - res)), | |||
There was a problem hiding this comment.
These math functions can kill your performance. You should benchmark these changes. Also might be worth comparing to sip hash using (const char*) &vertex->lat as input. I wonder which is faster.
There was a problem hiding this comment.
Even better would be to assign an index to the individual vertices, removing the need for floating point equality/hashing entirely.
There was a problem hiding this comment.
Agreed - I'd suggest we don't focus on the performance in this diff (can't imagine that a single fabs call will make that big a difference), and come back to it when the vertex index mode change is implemented. I'll open an issue for that now.
mrdvt92
pushed a commit
to mrdvt92/h3
that referenced
this pull request
Jun 19, 2022
Fix segfault in `removeVertexNode` due to negative hash
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.
Fixes #92
The hash in
_hashVertexproduces out of range values if(vertex->lat + vertex->lon)was negative, because it would produce a negative remainder (such as -1) and then cast it to
an unsigned integer type. This index would then cause a segfault.
testH3SetToLinkedGeo.c is also updated to not
calloca fixed size of 1.