[ 
https://issues.apache.org/jira/browse/GUACAMOLE-662?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Michael Jumper updated GUACAMOLE-662:
-------------------------------------
    Description: 
The unit test for the "nest" instruction (hopefully soon to be deprecated via 
GUACAMOLE-661) has been failing. Assuming the failure is legitimate, it hasn't 
been a problem in practice since the "nest" instruction has been unused for 
some time, but the fact that this failure has gone unnoticed is troubling. The 
build results currently show the following for {{make check}}:

{code:none}
...
PASS: test_libguac
============================================================================
Testsuite summary for guacamole-server 1.0.0
============================================================================
# TOTAL: 1
# PASS:  1
# SKIP:  0
# XFAIL: 0
# FAIL:  0
# XPASS: 0
# ERROR: 0
============================================================================
...
{code}

The test log within {{tests/test_libguac.log}} contradicts this, showing that 
there is actually a test failure (not to mention more than one test total):

{code:none}


     CUnit - A unit testing framework for C - Version 2.1-3
     http://cunit.sourceforge.net/


Suite: protocol
  Test: base64-decode ...passed
  Test: instruction-parse ...passed
  Test: instruction-read ...passed
  Test: instruction-write ...passed
  Test: nest-write ...FAILED
    1. protocol/nest_write.c:104  - CU_ASSERT_STRING_EQUAL(buffer,expected)
Suite: client
  Test: layer-pool ...passed
  Test: buffer-pool ...passed
Suite: util
  Test: guac-pool ...passed
  Test: guac-unicode ...passed

Run Summary:    Type  Total    Ran Passed Failed Inactive
              suites      3      3    n/a      0        0
               tests      9      9      8      1        0
             asserts  11091  11091  11090      1      n/a

Elapsed time =   -0.006 seconds
{code}

If "nest" is truly not behaving correctly, it should be corrected. If "nest" is 
correct but the test is wrong, the test should be fixed. Most importantly, the 
testing portion of the build process should be corrected such that:

# The test report actually captures the correct number of tests passing and 
test failures.
# A failing test fails the build.


  was:
The unit test for the "nest" instruction (hopefully soon to be deprecated via 
GUACAMOLE-611) has been failing. Assuming the failure is legitimate, it hasn't 
been a problem in practice since the "nest" instruction has been unused for 
some time, but the fact that this failure has gone unnoticed is troubling. The 
build results currently show the following for {{make check}}:

{code:none}
...
PASS: test_libguac
============================================================================
Testsuite summary for guacamole-server 1.0.0
============================================================================
# TOTAL: 1
# PASS:  1
# SKIP:  0
# XFAIL: 0
# FAIL:  0
# XPASS: 0
# ERROR: 0
============================================================================
...
{code}

The test log within {{tests/test_libguac.log}} contradicts this, showing that 
there is actually a test failure (not to mention more than one test total):

{code:none}


     CUnit - A unit testing framework for C - Version 2.1-3
     http://cunit.sourceforge.net/


Suite: protocol
  Test: base64-decode ...passed
  Test: instruction-parse ...passed
  Test: instruction-read ...passed
  Test: instruction-write ...passed
  Test: nest-write ...FAILED
    1. protocol/nest_write.c:104  - CU_ASSERT_STRING_EQUAL(buffer,expected)
Suite: client
  Test: layer-pool ...passed
  Test: buffer-pool ...passed
Suite: util
  Test: guac-pool ...passed
  Test: guac-unicode ...passed

Run Summary:    Type  Total    Ran Passed Failed Inactive
              suites      3      3    n/a      0        0
               tests      9      9      8      1        0
             asserts  11091  11091  11090      1      n/a

Elapsed time =   -0.006 seconds
{code}

If "nest" is truly not behaving correctly, it should be corrected. If "nest" is 
correct but the test is wrong, the test should be fixed. Most importantly, the 
testing portion of the build process should be corrected such that:

# The test report actually captures the correct number of tests passing and 
test failures.
# A failing test fails the build.



> Failing unit tests for guacamole-server not triggering build failure
> --------------------------------------------------------------------
>
>                 Key: GUACAMOLE-662
>                 URL: https://issues.apache.org/jira/browse/GUACAMOLE-662
>             Project: Guacamole
>          Issue Type: Bug
>          Components: guacamole-server
>            Reporter: Michael Jumper
>            Priority: Major
>
> The unit test for the "nest" instruction (hopefully soon to be deprecated via 
> GUACAMOLE-661) has been failing. Assuming the failure is legitimate, it 
> hasn't been a problem in practice since the "nest" instruction has been 
> unused for some time, but the fact that this failure has gone unnoticed is 
> troubling. The build results currently show the following for {{make check}}:
> {code:none}
> ...
> PASS: test_libguac
> ============================================================================
> Testsuite summary for guacamole-server 1.0.0
> ============================================================================
> # TOTAL: 1
> # PASS:  1
> # SKIP:  0
> # XFAIL: 0
> # FAIL:  0
> # XPASS: 0
> # ERROR: 0
> ============================================================================
> ...
> {code}
> The test log within {{tests/test_libguac.log}} contradicts this, showing that 
> there is actually a test failure (not to mention more than one test total):
> {code:none}
>      CUnit - A unit testing framework for C - Version 2.1-3
>      http://cunit.sourceforge.net/
> Suite: protocol
>   Test: base64-decode ...passed
>   Test: instruction-parse ...passed
>   Test: instruction-read ...passed
>   Test: instruction-write ...passed
>   Test: nest-write ...FAILED
>     1. protocol/nest_write.c:104  - CU_ASSERT_STRING_EQUAL(buffer,expected)
> Suite: client
>   Test: layer-pool ...passed
>   Test: buffer-pool ...passed
> Suite: util
>   Test: guac-pool ...passed
>   Test: guac-unicode ...passed
> Run Summary:    Type  Total    Ran Passed Failed Inactive
>               suites      3      3    n/a      0        0
>                tests      9      9      8      1        0
>              asserts  11091  11091  11090      1      n/a
> Elapsed time =   -0.006 seconds
> {code}
> If "nest" is truly not behaving correctly, it should be corrected. If "nest" 
> is correct but the test is wrong, the test should be fixed. Most importantly, 
> the testing portion of the build process should be corrected such that:
> # The test report actually captures the correct number of tests passing and 
> test failures.
> # A failing test fails the build.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to