Clone Tools
  • last updated a few minutes ago
Constraints: committers
Constraints: files
Constraints: dates
[BYTEMAN-328] handling dtest library instrumented classes in concurrent ready way

Adding tests against rule construction

tweaked a few calls to helper out/error/verbose -- part fixes for BYTEMAN-294

Merge pull request #49 from spolti/BYTEMAN-294

[BYTEMAN-294] - Improve internal trace logging

delete asciidoc module from docs pom so we can successfully build using jdk9

modify agent pom so Byteman agent uses ASM6.0_ALPHA -- fix for BYTEMAN-324

[BYTEMAN-294] - Improve internal trace logging

Updating byteman sample ThreadHistoryMonitorHelper to store all events

Updating and refactoring ThreadHistoryMonitorHelper to get stored all

events which happens during program execution.

Till now Map was used with key based on thread name. Thread name can be easily

the same for several threads. History was quite sparse.

By my understanding thread event history should be a log that stores all

events and it's up to developer to query the results.

That's the reson for this change.

    • -1
    • +11
Adding few instrumented class calls to get count info

rolled over pom versions to next snapshot

    • -1
    • +1
roll over pom versions for 3.0.6 release

    • -1
    • +1
Disable compilation when injecting into inner classes -- part-fix for BYTEMAN-288

References to inner classes from compiled rule code cause verify

errors. However, these errors can be avoided if instead the ruel code

is interpreted. This fix disables compilation of rule code for any

INTERFACE rule or overriding CLASS (with a ^) rule whose trigger class

is found to be an inner class. Note that these two cases are treated

specially because there are legitimate cases where compilation might

be appropriate for some trigger classes which match the rule but will

cause problems with other matches. If compilation is specified using a

normal (non-overriding) CLASS rule and the target class is an inner

class then that is a spec error so Byteman generates bytecode and an

error results.

This is merely a limited heuristic which has both false positives and

negatives. It can interpret rules whcih might legitimately compile

(e.g. they don't use $0) and pass rules which will still fail to

compile (e.g. $1 is an inner class). It also doesn't detect cases

where the inner class is non-static (guaranteed to fail if we try to

use the inner instance with no handle on the enclosing wrapper

instance). A better fix requires upgarding the type check methods of

Expression classes to detect uses of inner classes and either disable

compilation or throw a type exception.

    • -0
    • +86
correct problem running bmunit unit tests with new surefire plugin

Enable rules (name/text) to be added directly to the rule check. Upgraded surefire plugin due to build issues

Enable rules (name/text) to be added directly to the rule check. Upgraded surefire plugin due to build issues

Enable rules (name/text) to be added directly to the rule check. Upgraded surefire plugin due to build issues

Merge branch 'spolti-BYTEMAN-294'

revert bmcheck driver class to not use Helper API for output

handle arbitrary exceptions from type check when performing bmcheck

Merge changes for BYTEMAN-294 to use Helper methods for message output

dtest rulebuilder comment formatted with html tags

changing directory structure of dtest for follow maven standards and for being able to create test directory

rolled over versions to next snapshot release 3.0.6-SNAPSHOT

    • -1
    • +1
rolled version round to 3.0.5 prior to release

    • -1
    • +1
Switch to asciidoc and add misssing content - completes fix for BYTEMAN-289,BYTEMAN-313, BYTEMAN-237

changed all occurrences of new XXX to XXX.valueOf for boxed numeric types XXX

ensure AT EXCEPTION EXIT rules check their bindings -- fix for BYTEMAN-237

enhancing dtest Instrumentor to work with 'INTERFACE' in all its methods

corrected fork attributes in failsafe test specifications

    • -8
    • +16
fix for BYTEMAN-320

fix jdk9 javadoc generation issue