<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugs.webkit.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.4.1"
          urlbase="https://bugs.webkit.org/"
          
          maintainer="admin@webkit.org"
>

    <bug>
          <bug_id>187066</bug_id>
          
          <creation_ts>2018-06-26 15:21:13 -0700</creation_ts>
          <short_desc>REGRESSION: [macOS Sierra] TestWebKitAPI.WebKit.WebsiteDataStoreCustomPaths is a flaky failure</short_desc>
          <delta_ts>2018-07-17 16:34:44 -0700</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>WebKit</product>
          <component>Tools / Tests</component>
          <version>WebKit Nightly Build</version>
          <rep_platform>Unspecified</rep_platform>
          <op_sys>Unspecified</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords>InRadar</keywords>
          <priority>P2</priority>
          <bug_severity>Normal</bug_severity>
          <target_milestone>---</target_milestone>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter name="Dawei Fenton (:realdawei)">realdawei</reporter>
          <assigned_to name="Sihui Liu">sihui_liu</assigned_to>
          <cc>beidson</cc>
    
    <cc>cdumez</cc>
    
    <cc>commit-queue</cc>
    
    <cc>fred.wang</cc>
    
    <cc>ggaren</cc>
    
    <cc>lforschler</cc>
    
    <cc>ryanhaddad</cc>
    
    <cc>sihui_liu</cc>
    
    <cc>tsavell</cc>
    
    <cc>webkit-bug-importer</cc>
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>1436688</commentid>
    <comment_count>0</comment_count>
    <who name="Dawei Fenton (:realdawei)">realdawei</who>
    <bug_when>2018-06-26 15:21:13 -0700</bug_when>
    <thetext>TestWebKitAPI.WebKit.WebsiteDataStoreCustomPaths is flaky failure (though frequent) on mac Sierra, I haven&apos;t been able to repro locally yet, but I&apos;m looking further into it.

Here are results from a recent run:
https://build.webkit.org/builders/Apple%20Sierra%20Release%20WK2%20%28Tests%29/builds/10169/steps/run-api-tests/logs/stdio

Failed

    TestWebKitAPI.WebKit.WebsiteDataStoreCustomPaths
        
        /Volumes/Data/slave/sierra-release/build/Tools/TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:187
        Value of: [[NSFileManager defaultManager] fileExistsAtPath:fileIDBPath.get().path]
          Actual: true
        Expected: false</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1437341</commentid>
    <comment_count>1</comment_count>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-06-28 08:28:49 -0700</bug_when>
    <thetext>Looks pretty bad, we ask the data store to remove all the IDB data, and wait for its completion handler. Then the IDB database path still exists.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1437634</commentid>
    <comment_count>2</comment_count>
    <who name="Radar WebKit Bug Importer">webkit-bug-importer</who>
    <bug_when>2018-06-28 17:02:12 -0700</bug_when>
    <thetext>&lt;rdar://problem/41609065&gt;</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1438521</commentid>
    <comment_count>3</comment_count>
    <who name="Ryan Haddad">ryanhaddad</who>
    <bug_when>2018-07-02 10:10:55 -0700</bug_when>
    <thetext>This test is frequently making the Sierra bots red. If we aren&apos;t going to fix it soon, can we disable it?</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1438542</commentid>
    <comment_count>4</comment_count>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-02 10:40:05 -0700</bug_when>
    <thetext>(In reply to Ryan Haddad from comment #3)
&gt; This test is frequently making the Sierra bots red. If we aren&apos;t going to
&gt; fix it soon, can we disable it?

Brady should comment.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1441669</commentid>
    <comment_count>5</comment_count>
    <who name="Ryan Haddad">ryanhaddad</who>
    <bug_when>2018-07-12 14:28:12 -0700</bug_when>
    <thetext>(In reply to Chris Dumez from comment #4)
&gt; (In reply to Ryan Haddad from comment #3)
&gt; &gt; This test is frequently making the Sierra bots red. If we aren&apos;t going to
&gt; &gt; fix it soon, can we disable it?
&gt; 
&gt; Brady should comment.
Brady, any thoughts here?</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442440</commentid>
    <comment_count>6</comment_count>
      <attachid>345103</attachid>
    <who name="Sihui Liu">sihui_liu</who>
    <bug_when>2018-07-16 11:34:10 -0700</bug_when>
    <thetext>Created attachment 345103
Patch</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442699</commentid>
    <comment_count>7</comment_count>
    <who name="Brady Eidson">beidson</who>
    <bug_when>2018-07-17 10:03:43 -0700</bug_when>
    <thetext>I think we need to know *why* this patch fixes the flakiness.

It appears to me that if this patch fixes the flakiness then there&apos;s a real bug in WebKit that clients could be running up against.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442703</commentid>
    <comment_count>8</comment_count>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-17 10:13:28 -0700</bug_when>
    <thetext>(In reply to Brady Eidson from comment #7)
&gt; I think we need to know *why* this patch fixes the flakiness.
&gt; 
&gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; bug in WebKit that clients could be running up against.

If a Child process crashes while there are pending callbacks then those callbacks will be resolved upon crash.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442705</commentid>
    <comment_count>9</comment_count>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-17 10:18:45 -0700</bug_when>
    <thetext>(In reply to Chris Dumez from comment #8)
&gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; 
&gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; bug in WebKit that clients could be running up against.
&gt; 
&gt; If a Child process crashes while there are pending callbacks then those
&gt; callbacks will be resolved upon crash.

Nah, in this case we&apos;re terminating the WebProcess and the pending completion handlers would be for the StorageProcess.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442706</commentid>
    <comment_count>10</comment_count>
      <attachid>345103</attachid>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-17 10:19:59 -0700</bug_when>
    <thetext>Comment on attachment 345103
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=345103&amp;action=review

&gt; Tools/ChangeLog:9
&gt; +        Killing the web process actually led to creating another new process and reloading the page 

&quot;reloading the page&quot; &lt;- Note that WebKit does not do that. It is up to the client so if this is happening then it means the API test is setting a crash handler and calling reload.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442717</commentid>
    <comment_count>11</comment_count>
    <who name="Brady Eidson">beidson</who>
    <bug_when>2018-07-17 10:52:22 -0700</bug_when>
    <thetext>(In reply to Chris Dumez from comment #9)
&gt; (In reply to Chris Dumez from comment #8)
&gt; &gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; &gt; 
&gt; &gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; &gt; bug in WebKit that clients could be running up against.
&gt; &gt; 
&gt; &gt; If a Child process crashes while there are pending callbacks then those
&gt; &gt; callbacks will be resolved upon crash.
&gt; 
&gt; Nah, in this case we&apos;re terminating the WebProcess and the pending
&gt; completion handlers would be for the StorageProcess.

Actually, you were right.

IDB goes to both Storage and Web processes (because private browsing IDB is in the web processes)</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442719</commentid>
    <comment_count>12</comment_count>
    <who name="Brady Eidson">beidson</who>
    <bug_when>2018-07-17 10:55:15 -0700</bug_when>
    <thetext>(In reply to Brady Eidson from comment #11)
&gt; (In reply to Chris Dumez from comment #9)
&gt; &gt; (In reply to Chris Dumez from comment #8)
&gt; &gt; &gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; &gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; &gt; &gt; bug in WebKit that clients could be running up against.
&gt; &gt; &gt; 
&gt; &gt; &gt; If a Child process crashes while there are pending callbacks then those
&gt; &gt; &gt; callbacks will be resolved upon crash.
&gt; &gt; 
&gt; &gt; Nah, in this case we&apos;re terminating the WebProcess and the pending
&gt; &gt; completion handlers would be for the StorageProcess.
&gt; 
&gt; Actually, you were right.
&gt; 
&gt; IDB goes to both Storage and Web processes (because private browsing IDB is
&gt; in the web processes)

Wait. Actually looking at the test, there&apos;s no callbacks here, though... are there?

What callbacks are you talking about here?

The test waits for activity from the web process then makes sure files are in place afterwards. That&apos;s all.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442764</commentid>
    <comment_count>13</comment_count>
    <who name="Sihui Liu">sihui_liu</who>
    <bug_when>2018-07-17 13:15:34 -0700</bug_when>
    <thetext>(In reply to Brady Eidson from comment #12)
&gt; (In reply to Brady Eidson from comment #11)
&gt; &gt; (In reply to Chris Dumez from comment #9)
&gt; &gt; &gt; (In reply to Chris Dumez from comment #8)
&gt; &gt; &gt; &gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; &gt; &gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; &gt; &gt; &gt; bug in WebKit that clients could be running up against.
&gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; If a Child process crashes while there are pending callbacks then those
&gt; &gt; &gt; &gt; callbacks will be resolved upon crash.
&gt; &gt; &gt; 
&gt; &gt; &gt; Nah, in this case we&apos;re terminating the WebProcess and the pending
&gt; &gt; &gt; completion handlers would be for the StorageProcess.
&gt; &gt; 
&gt; &gt; Actually, you were right.
&gt; &gt; 
&gt; &gt; IDB goes to both Storage and Web processes (because private browsing IDB is
&gt; &gt; in the web processes)
&gt; 
&gt; Wait. Actually looking at the test, there&apos;s no callbacks here, though... are
&gt; there?
&gt; 
&gt; What callbacks are you talking about here?
&gt; 
&gt; The test waits for activity from the web process then makes sure files are
&gt; in place afterwards. That&apos;s all.

My understanding for this problem is: when reloading the page, the messages from the HTML file would be sent again and the message handler kept setting flag receivedScriptMessage true; so we didn&apos;t wait for completion handler to finish to continue. But the assertion is supposed to be reached and evaluated after the completion handler finishes (which means the deletion finishes).</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442765</commentid>
    <comment_count>14</comment_count>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-17 13:20:32 -0700</bug_when>
    <thetext>(In reply to Sihui Liu from comment #13)
&gt; (In reply to Brady Eidson from comment #12)
&gt; &gt; (In reply to Brady Eidson from comment #11)
&gt; &gt; &gt; (In reply to Chris Dumez from comment #9)
&gt; &gt; &gt; &gt; (In reply to Chris Dumez from comment #8)
&gt; &gt; &gt; &gt; &gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; &gt; &gt; &gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; &gt; &gt; &gt; &gt; bug in WebKit that clients could be running up against.
&gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; If a Child process crashes while there are pending callbacks then those
&gt; &gt; &gt; &gt; &gt; callbacks will be resolved upon crash.
&gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; Nah, in this case we&apos;re terminating the WebProcess and the pending
&gt; &gt; &gt; &gt; completion handlers would be for the StorageProcess.
&gt; &gt; &gt; 
&gt; &gt; &gt; Actually, you were right.
&gt; &gt; &gt; 
&gt; &gt; &gt; IDB goes to both Storage and Web processes (because private browsing IDB is
&gt; &gt; &gt; in the web processes)
&gt; &gt; 
&gt; &gt; Wait. Actually looking at the test, there&apos;s no callbacks here, though... are
&gt; &gt; there?
&gt; &gt; 
&gt; &gt; What callbacks are you talking about here?
&gt; &gt; 
&gt; &gt; The test waits for activity from the web process then makes sure files are
&gt; &gt; in place afterwards. That&apos;s all.
&gt; 
&gt; My understanding for this problem is: when reloading the page, the messages
&gt; from the HTML file would be sent again and the message handler kept setting
&gt; flag receivedScriptMessage true; so we didn&apos;t wait for completion handler to
&gt; finish to continue. But the assertion is supposed to be reached and
&gt; evaluated after the completion handler finishes (which means the deletion
&gt; finishes).

Who reloads the page? The API test does not seem to register a crash handler.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442780</commentid>
    <comment_count>15</comment_count>
    <who name="Brady Eidson">beidson</who>
    <bug_when>2018-07-17 13:58:50 -0700</bug_when>
    <thetext>(In reply to Chris Dumez from comment #14)
&gt; (In reply to Sihui Liu from comment #13)
&gt; &gt; (In reply to Brady Eidson from comment #12)
&gt; &gt; &gt; (In reply to Brady Eidson from comment #11)
&gt; &gt; &gt; &gt; (In reply to Chris Dumez from comment #9)
&gt; &gt; &gt; &gt; &gt; (In reply to Chris Dumez from comment #8)
&gt; &gt; &gt; &gt; &gt; &gt; (In reply to Brady Eidson from comment #7)
&gt; &gt; &gt; &gt; &gt; &gt; &gt; I think we need to know *why* this patch fixes the flakiness.
&gt; &gt; &gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; &gt; &gt; It appears to me that if this patch fixes the flakiness then there&apos;s a real
&gt; &gt; &gt; &gt; &gt; &gt; &gt; bug in WebKit that clients could be running up against.
&gt; &gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; &gt; If a Child process crashes while there are pending callbacks then those
&gt; &gt; &gt; &gt; &gt; &gt; callbacks will be resolved upon crash.
&gt; &gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; &gt; Nah, in this case we&apos;re terminating the WebProcess and the pending
&gt; &gt; &gt; &gt; &gt; completion handlers would be for the StorageProcess.
&gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; Actually, you were right.
&gt; &gt; &gt; &gt; 
&gt; &gt; &gt; &gt; IDB goes to both Storage and Web processes (because private browsing IDB is
&gt; &gt; &gt; &gt; in the web processes)
&gt; &gt; &gt; 
&gt; &gt; &gt; Wait. Actually looking at the test, there&apos;s no callbacks here, though... are
&gt; &gt; &gt; there?
&gt; &gt; &gt; 
&gt; &gt; &gt; What callbacks are you talking about here?
&gt; &gt; &gt; 
&gt; &gt; &gt; The test waits for activity from the web process then makes sure files are
&gt; &gt; &gt; in place afterwards. That&apos;s all.
&gt; &gt; 
&gt; &gt; My understanding for this problem is: when reloading the page, the messages
&gt; &gt; from the HTML file would be sent again and the message handler kept setting
&gt; &gt; flag receivedScriptMessage true; so we didn&apos;t wait for completion handler to
&gt; &gt; finish to continue. But the assertion is supposed to be reached and
&gt; &gt; evaluated after the completion handler finishes (which means the deletion
&gt; &gt; finishes).
&gt; 
&gt; Who reloads the page? The API test does not seem to register a crash handler.

Just looked at the backtrace on Sihui&apos;s computer. It&apos;s definitely built in to WebKit, and is WebKit&apos;s default behavior when there is no crash handler.

The solution seems to be to install a crash handler. That would prevent WebKit&apos;s automatic reloading.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442817</commentid>
    <comment_count>16</comment_count>
      <attachid>345199</attachid>
    <who name="Sihui Liu">sihui_liu</who>
    <bug_when>2018-07-17 15:46:33 -0700</bug_when>
    <thetext>Created attachment 345199
Patch</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442820</commentid>
    <comment_count>17</comment_count>
      <attachid>345199</attachid>
    <who name="Chris Dumez">cdumez</who>
    <bug_when>2018-07-17 15:48:07 -0700</bug_when>
    <thetext>Comment on attachment 345199
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=345199&amp;action=review

&gt; Tools/TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:61
&gt; +    // Overwrite thte default policy which launches a new web process and reload page on crash.

typo: thte
s/reload/reloads</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442822</commentid>
    <comment_count>18</comment_count>
      <attachid>345201</attachid>
    <who name="Sihui Liu">sihui_liu</who>
    <bug_when>2018-07-17 15:54:09 -0700</bug_when>
    <thetext>Created attachment 345201
Patch for landing</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442844</commentid>
    <comment_count>19</comment_count>
      <attachid>345201</attachid>
    <who name="WebKit Commit Bot">commit-queue</who>
    <bug_when>2018-07-17 16:34:43 -0700</bug_when>
    <thetext>Comment on attachment 345201
Patch for landing

Clearing flags on attachment: 345201

Committed r233899: &lt;https://trac.webkit.org/changeset/233899&gt;</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1442845</commentid>
    <comment_count>20</comment_count>
    <who name="WebKit Commit Bot">commit-queue</who>
    <bug_when>2018-07-17 16:34:44 -0700</bug_when>
    <thetext>All reviewed patches have been landed.  Closing bug.</thetext>
  </long_desc>
      
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>345103</attachid>
            <date>2018-07-16 11:34:10 -0700</date>
            <delta_ts>2018-07-17 15:46:31 -0700</delta_ts>
            <desc>Patch</desc>
            <filename>bug-187066-20180716113409.patch</filename>
            <type>text/plain</type>
            <size>2013</size>
            <attacher name="Sihui Liu">sihui_liu</attacher>
            
              <data encoding="base64">U3VidmVyc2lvbiBSZXZpc2lvbjogMjMzODQ5CmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg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</data>

          </attachment>
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>345199</attachid>
            <date>2018-07-17 15:46:33 -0700</date>
            <delta_ts>2018-07-17 15:54:06 -0700</delta_ts>
            <desc>Patch</desc>
            <filename>bug-187066-20180717154632.patch</filename>
            <type>text/plain</type>
            <size>2966</size>
            <attacher name="Sihui Liu">sihui_liu</attacher>
            
              <data encoding="base64">U3VidmVyc2lvbiBSZXZpc2lvbjogMjMzODY4CmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg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</data>

          </attachment>
          <attachment
              isobsolete="0"
              ispatch="1"
              isprivate="0"
          >
            <attachid>345201</attachid>
            <date>2018-07-17 15:54:09 -0700</date>
            <delta_ts>2018-07-17 16:34:43 -0700</delta_ts>
            <desc>Patch for landing</desc>
            <filename>bug-187066-20180717155408.patch</filename>
            <type>text/plain</type>
            <size>2962</size>
            <attacher name="Sihui Liu">sihui_liu</attacher>
            
              <data encoding="base64">U3VidmVyc2lvbiBSZXZpc2lvbjogMjMzODY4CmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg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==
</data>

          </attachment>
      

    </bug>

</bugzilla>