John Cortell
|
1ee7bff8d5
|
Added test for interrupt
|
2010-06-01 19:18:33 +00:00 |
|
John Cortell
|
ae33c729a7
|
Missing asterisk in javadoc comment
|
2010-06-01 19:02:34 +00:00 |
|
John Cortell
|
f001fc32d0
|
Fixed intermittent test failure caused by buggy test. A breakpoint hit can arrive after SyncUtil.resumeUntilStopped() returns. The test needs to call waitForBreakpointEvent() afterwards.
|
2010-06-01 18:44:26 +00:00 |
|
John Cortell
|
429c116b18
|
Stop trying to side-step the fact that on Windows interrupting the target can leave it suspended prematurely, only to have it suspend again when you resume it. My previous attempt worked only some of the time. Trying to close the gap would be way too messy, so just skip the "while target running" tests on Windows.
|
2010-06-01 18:07:55 +00:00 |
|
John Cortell
|
0874a4a973
|
Don't wait forever
|
2010-05-28 20:49:52 +00:00 |
|
John Cortell
|
18f37c3ccf
|
Bug 314174: Eliminate the 1s wait at start of each DSF-GDB test. (Avoid possible NPE when a test fails)
|
2010-05-28 19:12:27 +00:00 |
|
Marc Khouzam
|
69444100a5
|
Bug 314628: Always restore the target state even if a failure has happened.
|
2010-05-28 18:48:01 +00:00 |
|
John Cortell
|
ca9f612e04
|
Incorrect usage of the special wait function in two places.
|
2010-05-27 22:56:38 +00:00 |
|
John Cortell
|
0ad079da76
|
Memory tests were failing with gdb 7.0. Marc's hunch was right. Use of UNIQUE_PROCESS_ID was the problem.
|
2010-05-27 22:31:26 +00:00 |
|
John Cortell
|
17f248d072
|
Bug 314174: Eliminate the 1s wait at start of each DSF-GDB test.
|
2010-05-27 22:11:53 +00:00 |
|
John Cortell
|
4972e64775
|
Adjust to the new dual POSIX/MSVC buildable MultiThread.cc. Also fix problems in the registers test
|
2010-05-27 21:20:45 +00:00 |
|
John Cortell
|
87aa0351e6
|
Bug 314174: Eliminate the 1s wait at start of each DSF-GDB test.
|
2010-05-27 21:08:28 +00:00 |
|
John Cortell
|
abef64cb07
|
Make dsf-gdb test programs buildable with MinGW. MinGW gdb has problems debugging a cygwin-built executable, so we need to build the test apps with MinGW when testing that debugger. Confirmed these modified files now build with both Cygwin and MinGW.
|
2010-05-26 19:38:11 +00:00 |
|
John Cortell
|
af21820551
|
Various expression tests where failing on Windows because of subtle differences in the compiler and the line tables it generates, and perhaps scoping behaviors. The code would end up on the trailing brace of a function, and all variables are out of scope there.
|
2010-05-25 23:01:40 +00:00 |
|
John Cortell
|
da1f29806f
|
Don't use the junit timeout feature. It's too rigid and is a pain when troubleshooting issues.
|
2010-05-25 22:47:12 +00:00 |
|
John Cortell
|
85162ebd30
|
Don't use the junit timeout feature. It's too rigid and is a pain when troubleshooting issues.
|
2010-05-25 22:45:29 +00:00 |
|
Marc Khouzam
|
30849ec0ce
|
Better use of container context
|
2010-05-12 13:23:44 +00:00 |
|
Marc Khouzam
|
9b222a1f27
|
Better error printout
|
2010-04-29 18:35:34 +00:00 |
|
Marc Khouzam
|
39668b20a5
|
Small enhancement to get the containerDmc in the unit tests
|
2010-04-29 14:21:47 +00:00 |
|
Marc Khouzam
|
0bd05a8dbc
|
Bug 310565: Support for new format for Apple GDB.
|
2010-04-27 19:08:40 +00:00 |
|
John Cortell
|
e776528588
|
[306427] Remove obsolete catchpoint types from CDT
|
2010-04-12 19:01:24 +00:00 |
|
John Cortell
|
92f88c096c
|
Centralize new method to get heavyweight process context
|
2010-04-12 15:57:59 +00:00 |
|
John Cortell
|
da04759802
|
Neglected to spin off the service call in getProcessContext() on the dsf executor
|
2010-04-09 22:55:26 +00:00 |
|
John Cortell
|
cbdc8d81d3
|
For the stop on main breakpoint (a temporaro breakpoint) gdb 7.0 reports the stopped event with a reason=breakpoint. It provides no reason in gdb < 7.0.
|
2010-04-09 22:44:45 +00:00 |
|
John Cortell
|
ed5b124ce7
|
Use run control service to get the actual process context. Fixes tests with MinGW 7.0. Also, had to adjust two tests to not assume threads are reported in a particular order
|
2010-04-09 18:05:08 +00:00 |
|
John Cortell
|
119d848c62
|
[305752] Moved gdb event breakpoint constants into new common gdb plugin as well as the registering of the event types (via an extension)
|
2010-04-08 20:02:35 +00:00 |
|
John Cortell
|
7c92312b57
|
minor. wrong cathcpoint type specified, but catchpoint creation is supposed to fail in this case anyway.
|
2010-04-08 14:53:17 +00:00 |
|
John Cortell
|
080bcc692d
|
Corrected argument order; error message is misleading, otherwise.
|
2010-04-07 19:49:00 +00:00 |
|
Marc Khouzam
|
8a2cae3b45
|
[248587] Added new catchpoint JUnit tests to test suites.
|
2010-04-07 18:39:56 +00:00 |
|
John Cortell
|
367979ea80
|
[248587] Support "Event' Breakpoints
|
2010-04-06 20:01:42 +00:00 |
|
John Cortell
|
57c6ba5fec
|
[248587] Support "Event' Breakpoints
|
2010-04-06 19:32:14 +00:00 |
|
John Cortell
|
4ec5b596be
|
Specify dwarf2 debug info. Default for 3.x is STABS, and that is an inferior format. At a minimum, C++ exception catchpoints don't work, and there is a class that tests them.
|
2010-04-05 17:38:41 +00:00 |
|
John Cortell
|
983dc28f18
|
Fixed inappropriate reference comparison. Should be a deep comparison (using equals method).
|
2010-03-31 20:00:58 +00:00 |
|
John Cortell
|
8b280845e4
|
Test case had coding goof. Incremented and checked wrong iterator variable, leading to early termination of the test.
|
2010-03-31 19:45:52 +00:00 |
|
John Cortell
|
56a2799dc0
|
javadoc enhancement
|
2010-03-29 16:00:01 +00:00 |
|
John Cortell
|
dc68630fb6
|
Fixed remaining edge case floating point math failure
|
2010-03-11 13:04:42 +00:00 |
|
John Cortell
|
c29c05bc25
|
Someone reported failures in the expressions tests because of small variations on floating point arithmetic results. E.g., expected '33.333333333333336" but got '33.333333333333329'". Floating point math is not exact, and we need to not be too precise in our expectations for those cases.
|
2010-03-10 21:38:34 +00:00 |
|
John Cortell
|
8087689bea
|
Fixed common programming goof
|
2010-03-10 20:15:26 +00:00 |
|
uid8941
|
d4367d5bba
|
Add -m32 flag to make the tests programs build properly on 64-bit machines.
|
2010-03-09 14:35:32 +00:00 |
|
John Cortell
|
0eb1964dd6
|
Don't wait forever
|
2010-03-05 23:24:17 +00:00 |
|
John Cortell
|
81a56f94b8
|
Fixed processes test. Shouldn't create a process context based on PID=UNIQUE_GROUP_ID. Also, relax the test for the process name. With 7.0 we get the path information in the name; in 6.8 we don't. Just check that the executable simple file name is in there.
|
2010-03-05 23:10:15 +00:00 |
|
John Cortell
|
0a695a3e81
|
Reworked awkward junit assertion flow (have seen this in various places)
|
2010-03-05 20:40:25 +00:00 |
|
John Cortell
|
428ad8d1f9
|
Additional cygwin-only check
|
2010-03-05 17:48:22 +00:00 |
|
John Cortell
|
4d3602121f
|
Fix tests for cygwin
|
2010-03-04 22:57:04 +00:00 |
|
John Cortell
|
ebdc876c9b
|
Fixed race condition test failure. The secondary thread, because it does nothing but a quick printf, was terminating before the test was able to check for its existence via IRunControl.getExecutionContexts(). Consequently, it would get one less thread than expected.
|
2010-03-04 22:14:37 +00:00 |
|
John Cortell
|
d84ca549b2
|
Write L-value test works with gdb >= 6.8, so refactor to make it available for those classes.
|
2010-03-04 17:17:44 +00:00 |
|
John Cortell
|
e6794e5a2e
|
fixed warnings
|
2010-03-04 16:48:01 +00:00 |
|
John Cortell
|
7eb06f6213
|
Fix runcontrol test on Windows. If the program is built with cygwin, there is an additional thread.
|
2010-03-04 16:43:00 +00:00 |
|
Marc Khouzam
|
20baa38a22
|
[304588] Return the problem number of children even if GDB counts private/public/protected as a level of children.
|
2010-03-04 15:48:37 +00:00 |
|
John Cortell
|
f642b20a88
|
fix indentation; no functional change
|
2010-03-04 15:30:35 +00:00 |
|