Nikolay Martynov
f71192266c
Rework the way multi-jvm tests work
...
To allow some JVMs runs only subset of tests
2018-10-02 13:20:19 -04:00
Nikolay Martynov
a4cded9b4f
Add some tests for distributed tracing in Tomcat and Jetty
2018-08-18 13:38:06 -04:00
Tyler Benson
3fc2738179
Improve test coverage and reduce duplication
2018-08-17 10:36:45 +10:00
Tyler Benson
0b6c1bc3d3
Fix tests for Java 7, 9, and 10
2018-08-16 14:23:58 +10:00
Tyler Benson
20993e7e16
Allow tests to run on all JDKs
2018-08-16 14:23:58 +10:00
Tyler Benson
e07dfffe5f
Migrate tests to new TestHttpServer
2018-08-16 14:23:58 +10:00
Tyler Benson
3f2ec7dc48
Move to separate packages
2018-08-16 12:24:27 +10:00
Tyler Benson
84d38a8270
Merge pull request #439 from DataDog/tyler/method-signatures
...
Cleanup method signatures and servlet 2 instrumentation
2018-08-13 17:29:02 +10:00
Tyler Benson
f801cbd466
Merge pull request #428 from DataDog/tyler/use-WeakConcurrentMap
...
Use WeakConcurrentMap
2018-08-13 16:44:16 +10:00
Tyler Benson
4583d12b31
Cleanup method signatures and servlet 2 instrumentation
...
adding the abstract servlet instrumentation reduces code duplication and makes it more similar to servlet 3 instrumentation.
2018-08-13 15:14:10 +10:00
Tyler Benson
eacc2d6402
Restructure how WeakMap is implemented
...
Use an interface.
Use a provider/supplier to avoid loading more onto the bootstrap classpath.
2018-08-08 16:54:44 +10:00
Tyler Benson
847484cd47
Use WeakConcurrentMap
...
Instead of `Collections.synchronizedMap(new WeakHashMap<>())`.
2018-08-06 14:14:57 +10:00
Tyler Benson
e7aa7c52b4
Improve comments and add synchronization.
2018-08-06 11:38:28 +10:00
Tyler Benson
064ae4c238
Separate span for input and output streams.
2018-08-03 15:41:16 +10:00
Nikolay Martynov
c66bd24d3a
Use `safeHasSuperType` instead of `isSubType`
...
`isSubType` may fail on certain class lookup problems, even on classes
unrelated to given instrumentation, preventing instrumentation from
being applied.
2018-07-31 20:14:03 -04:00
Andrew Kent
ee594f8567
Use connect operationName in httpurlconnection connect-first case
2018-07-26 15:34:27 -07:00
Andrew Kent
54485c3296
Don't trace httpurlconnection as a top-level trace
2018-07-26 15:34:27 -07:00
Andrew Kent
d598515d09
Handle cases where connect() is called first
2018-07-26 15:34:27 -07:00
Andrew Kent
df95af53a7
Remove getResponseCode() from httpurlconneciton instrumentation
2018-07-26 15:34:27 -07:00
Andrew Kent
2a35ecb497
Use connected field to determine io status of httpurlconnection
2018-07-26 15:34:27 -07:00
Gary Huang
2780ac7ac1
POC sun JDK http urlconnection instrumentation, currently creates one span.
2018-07-26 15:34:27 -07:00
Nikolay Martynov
d4d770fe42
Pull out ratpack helper into seprate class to avoid copy-paste
2018-07-23 10:04:53 -04:00
Andrew Kent
02a3e6a5d3
New Instrumentation API
2018-07-06 17:17:32 -04:00
Nikolay Martynov
60a0722390
Increase timeouts in UrlConnection test
...
Before those timeouts where set to 10ms which legitimatelly can come
before server (even localhost) has a chance to reply with 'connection
refused'.
This should fix flaky test.
2018-06-22 16:57:31 -04:00
Tyler Benson
c2771753b7
Suggested changes.
2018-06-08 16:38:51 +10:00
Nikolay Martynov
28c7c30af4
[HttpUrlConnection] Fix HEAD responses
...
Explain to ratpack that sending body for HEAD requests is not up to
spec [1] and seems to be confusing to clients. This resolves 'Keep-Alive'
mistery and makes disabling 'keepa-live' unnecessary.
[1] https://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html
2018-06-07 12:24:16 -04:00
Tyler Benson
9ff09b93e1
Refactor HttpUrlConnection Instrumentation
...
getOutputStream shouldn’t call getResponseCode, otherwise the request gets screwed up.
2018-06-07 16:25:38 +10:00