Changeset 231938 in webkit


Ignore:
Timestamp:
May 17, 2018 6:41:58 PM (6 years ago)
Author:
sbarati@apple.com
Message:

Remove shrinkFootprint test from apitests since it's flaky
https://bugs.webkit.org/show_bug.cgi?id=185754

Reviewed by Mark Lam.

This test is flaky as it keeps failing on certain people's machines.
Having a test about OS footprint seems like it'll forever be doomed
to being flaky.

  • API/tests/testapi.mm:

(testObjectiveCAPIMain):

Location:
trunk/Source/JavaScriptCore
Files:
2 edited

Legend:

Unmodified
Added
Removed
  • trunk/Source/JavaScriptCore/API/tests/testapi.mm

    r231589 r231938  
    14731473    }
    14741474
    1475     @autoreleasepool {
    1476         JSContext *context = [[JSContext alloc] init];
    1477         JSVirtualMachine *vm = [context virtualMachine];
    1478         [vm shrinkFootprint]; // Make sure that when we allocate a ton of memory below we reuse at little as possible.
    1479 
    1480         std::optional<size_t> footprintBefore = WTF::memoryFootprint();
    1481         RELEASE_ASSERT(footprintBefore);
    1482 
    1483         [context evaluateScript:@"for (let i = 0; i < 10000; ++i) { eval(`myVariable_${i} = [i]`); }"];
    1484 
    1485         static constexpr size_t approximateBytes = 10000 * sizeof(int);
    1486         std::optional<size_t> footprintMiddle = WTF::memoryFootprint();
    1487         RELEASE_ASSERT(footprintMiddle);
    1488         checkResult(@"Footprint is larger than what we allocated", *footprintMiddle > approximateBytes);
    1489         checkResult(@"Footprint got larger as we allocated a ton of stuff", *footprintMiddle > *footprintBefore);
    1490         size_t allocationDelta = *footprintMiddle - *footprintBefore;
    1491         checkResult(@"We allocated as much or more memory than what we expected to", allocationDelta >= approximateBytes);
    1492 
    1493         [context evaluateScript:@"for (let i = 0; i < 10000; ++i) { eval(`myVariable_${i} = null`); }"];
    1494         [vm shrinkFootprint];
    1495         std::optional<size_t> footprintAfter = WTF::memoryFootprint();
    1496         RELEASE_ASSERT(footprintAfter);
    1497         checkResult(@"Footprint got smaller after we shrank the VM", *footprintAfter < *footprintMiddle);
    1498         size_t freeDelta = *footprintMiddle - *footprintAfter;
    1499         checkResult(@"Shrinking the footprint of the VM actually freed memory", freeDelta > (approximateBytes / 2));
    1500     }
    1501 
    15021475    currentThisInsideBlockGetterTest();
    15031476    runDateTests();
  • trunk/Source/JavaScriptCore/ChangeLog

    r231930 r231938  
     12018-05-17  Saam Barati  <sbarati@apple.com>
     2
     3        Remove shrinkFootprint test from apitests since it's flaky
     4        https://bugs.webkit.org/show_bug.cgi?id=185754
     5
     6        Reviewed by Mark Lam.
     7
     8        This test is flaky as it keeps failing on certain people's machines.
     9        Having a test about OS footprint seems like it'll forever be doomed
     10        to being flaky.
     11
     12        * API/tests/testapi.mm:
     13        (testObjectiveCAPIMain):
     14
    1152018-05-17  Saam Barati  <sbarati@apple.com>
    216
Note: See TracChangeset for help on using the changeset viewer.