Merge "logcat: test: increase retry on fgetLongTimestamp to 4, and sleep to settle to 3" (6d1cabbf) · Commits · e / os / android_system_core

logcat/tests/logcat_test.cpp

+10 −7

Original line number	Diff line number	Diff line
		@@ -171,8 +171,8 @@ char fgetLongTime(char buffer, size_t buflen, FILE *fp) {
		if (!ep) {
		continue;
		}
		static const size_t tag_field_width = 7;
		ep -= tag_field_width;
		static const size_t pid_field_width = 7;
		ep -= pid_field_width;
		*ep = '\0';
		return cp;
		}
		@@ -184,14 +184,15 @@ char fgetLongTime(char buffer, size_t buflen, FILE *fp) {
		static size_t inject(ssize_t count) {
		if (count <= 0) return 0;

		static const size_t retry = 3;
		static const size_t retry = 4;
		size_t errors = retry;
		size_t num = 0;
		for(;;) {
		log_time ts(CLOCK_MONOTONIC);
		if (__android_log_btwrite(0, EVENT_TYPE_LONG, &ts, sizeof(ts)) >= 0) {
		if (++num >= (size_t)count) {
		sleep(1); // let data settle end-to-end
		// let data settle end-to-end
		sleep(3);
		return num;
		}
		errors = retry;
		@@ -211,7 +212,7 @@ TEST(logcat, tz) {
		return;
		}

		int tries = 3; // in case run too soon after system start or buffer clear
		int tries = 4; // in case run too soon after system start or buffer clear
		int count;

		do {
		@@ -228,6 +229,8 @@ TEST(logcat, tz) {
		while (fgetLongTime(buffer, sizeof(buffer), fp)) {
		if (strstr(buffer, " -0700") \|\| strstr(buffer, " -0800")) {
		++count;
		} else {
		fprintf(stderr, "ts=\"%s\"\n", buffer + 2);
		}
		}

		@@ -261,7 +264,7 @@ TEST(logcat, ntz) {
		}

		void do_tail(int num) {
		int tries = 3; // in case run too soon after system start or buffer clear
		int tries = 4; // in case run too soon after system start or buffer clear
		int count;

		do {
		@@ -310,7 +313,7 @@ TEST(logcat, tail_time) {
		char *first_timestamp = NULL;
		char *cp;

		int tries = 3; // in case run too soon after system start or buffer clear
		int tries = 4; // in case run too soon after system start or buffer clear

		// Do not be tempted to use -v usec because that increases the
		// chances of an occasional test failure by 1000 (see below).