-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 All,
Some interesting[1] reading: https://steveloughran.blogspot.com/2015/09/time-on-multi-core-multi-sock et-servers.html - -chris [1] For some definitions of "interesting" On 2/27/20 17:26, Christopher Schultz wrote: > Rémy, > > On 2/27/20 17:20, Christopher Schultz wrote: >> Rémy, > >> On 2/27/20 10:49, r...@apache.org wrote: >>> This is an automated email from the ASF dual-hosted git >>> repository. > >>> remm pushed a commit to branch master in repository >>> https://gitbox.apache.org/repos/asf/tomcat.git > > >>> The following commit(s) were added to refs/heads/master by >>> this push: new 22ad695 Update request start time using >>> nanoTime 22ad695 is described below > >>> commit 22ad69571c019f4d84ccc522298dddb4f8fa8d70 Author: remm >>> <r...@apache.org> AuthorDate: Thu Feb 27 16:49:04 2020 +0100 > >>> Update request start time using nanoTime > >>> get/setStartTime are still there, not sure about existing use. >>> Another patch round could deprecate them. Also change the >>> access log to be the start of the request (a small part of >>> 63286). --- > >>> [snip] > >>> diff --git a/java/org/apache/coyote/AbstractProcessor.java >>> b/java/org/apache/coyote/AbstractProcessor.java index >>> 254950e..5af3710 100644 --- >>> a/java/org/apache/coyote/AbstractProcessor.java +++ >>> b/java/org/apache/coyote/AbstractProcessor.java @@ -978,6 >>> +978,7 @@ public abstract class AbstractProcessor extends >>> AbstractProcessorLight implement >>> setSocketWrapper(socketWrapper); // Setup the minimal request >>> information request.setStartTime(System.currentTimeMillis()); >>> + request.setStartTimeNanos(System.nanoTime()); // Setup the >>> minimal response information response.setStatus(400); >>> response.setError(); > >> Since we are talking about nanoseconds, here, we ARE, by >> definition, splitting hairs :) > >> System.currentTimeMillis() and System.nanoTime() may disagree >> when you capture them like this, because [a super-small amount >> of] time elapses between the two calls. > >> Should we instead do: > >> long nanos = System.nanoTime(); >> request.setStartTimeNanos(nanos); request.setStartTime(nanos / >> 1000000l); > >> Or maybe: > >> long nanos = System.nanoTime(); >> request.setStartTimeNanos(nanos); >> request.setStartTime(TimeUnit.NANOSECONDS.toMillis(nanos)); > >> ? > > Also, System.nanoTime apparently takes a lot longer than > System.currentTimeMillis(), but it always moves forward in time, > while System.currentTimeMillis can fall back for leap seconds, etc. > and can behave in surprising ways :) > > -chris > -----BEGIN PGP SIGNATURE----- Comment: Using GnuPG with Thunderbird - https://www.enigmail.net/ iQIzBAEBCAAdFiEEMmKgYcQvxMe7tcJcHPApP6U8pFgFAl5YQz0ACgkQHPApP6U8 pFhE6A//aCZRXyBVu5VjqeM53Sa3luTca3+Q3o70cRbefrqSpgGHVGCaCEBTFX8L fIk1yrXE8TB9sA0C2d/SEdPK/Db0crirNTR6yQr49w5XU7a841u/wrI+Vqb3/5je +Jhg5yRyrSZD7LNyRmXk1xudjWehCGmixMQzZuRWTN1fJZsVKe1PUHvQ4omWkQ6u 8IiPfPyY3UNPw0bCofw291i/KSP23/Hdad+QGG/7HRdJR7XZqtIG7x9x9W27m12F 4kKjNfPvtDr/TAEjVvFw35L5G5hw4oAtk1pPbXabQItieU0iY78pIFJVE/PH4fQ9 Cjn62fPXtBXy/XI1WTYDTUkkLiu69vXzN9Za1kLXvH6rJuzo+0kkbEOudknwEbHz tTy5mGcdURv+qBloaH4PaednVeFFbVuU/vsZghh1ytNW379FvXtT3upuM+eJNPM1 US2v8TKSp3sK9kirgOWdwzAMkqXsQ7vQDGAiSZ+EMvMzn4CtOTWfaYioDvowzEIT thc3j/XmvhQtygQAgO7fG4cRVUEnKlJ9AhWmxzl6isNSA7zBMMsYBSivRKFeyOha 61jeUhbZzituUyGjzTxFJPPlsXY4+8XnuMSRGo68ZPlkAU8ZWtrrKO4iGZ51gY3L AN4ppdLGy6DeJM20nK21Xpjx31QW5qhcNGolL0llihWZW31hYLQ= =yi5w -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
