Andrew Johnson
ffcbd4ca8b
Merge changes from 3.14.12.5 (revno 12582)
2015-03-25 12:53:13 -05:00
Andrew Johnson
9470830091
catools: Fix printing -ve DBF_LONG values on 64-bit
2015-03-23 15:59:11 -05:00
Andrew Johnson
f498b36438
Prevent overflows in ao value conversion.
...
Also makes ROFF fields unsigned, needed for ROFF=0x80000000
2015-03-20 16:34:14 -05:00
Andrew Johnson
490540b005
Merged from 3.14 branch up to revno 12575
2015-03-19 10:22:15 -05:00
Andrew Johnson
eaef9aabdf
Fix export type of .LINK$ fields, add tests
2015-03-18 16:48:07 -05:00
Andrew Johnson
1fd91f52e2
dbStatic: Add field type lookup routines
2015-03-18 16:34:07 -05:00
Andrew Johnson
f994944531
libCom/posix: Remove pthread_setcanceltype()
2015-03-18 11:36:06 -05:00
Andrew Johnson
e1b2a0c5a3
Merged changes from 3.14 branch up to revno 12573
2015-03-13 11:50:26 -05:00
Andrew Johnson
5ffda58351
ca: Fix missing guard argument
2015-03-13 11:08:05 -05:00
Andrew Johnson
f6373093a5
Fix Windows linkage warning in alarmString.h
2015-03-13 10:56:03 -05:00
Andrew Johnson
8215244a28
Fix Windows time discontinuity report
2015-03-13 10:06:23 -05:00
Andrew Johnson
99f1dbb75e
Merged changes from 3.14 branch up to revno 12570
2015-03-11 16:26:15 -05:00
Andrew Johnson
de298da94c
Fix Jenkins instability on Windows
2015-03-11 10:53:37 -05:00
Andrew Johnson
753bfcc579
Check sender threads are starting
2015-03-10 18:14:11 -05:00
Andrew Johnson
4568cd3616
Chasing Jenkins instability on Windows
2015-03-10 14:17:28 -05:00
Andrew Johnson
1cacb8ef72
Make TOP an absolute path when running test programs
...
If the test starts an IOC, the registerRecordDeviceDriver
routine will compare the absolute path of TOP at build time
against the environment variable value (normally relative)
and print a warning if they differ.
This change cleans up the .tap output from such tests.
2015-03-10 13:04:15 -05:00
Michael Davidsaver
fb7273046c
epicsAtomic: also prevent vxWorks from using posix/epicsAtomicOSD.cpp
2015-03-10 10:39:37 -04:00
Michael Davidsaver
5bbd954ec8
epicsAtomic: fix noop OSDs
...
Ensure that all are epicsAtomicOSD.cpp and use
no-op to prevent RTEMS from using the posix
implementation.
2015-03-09 21:27:49 -04:00
Michael Davidsaver
6bf3f5554f
epicsAtomic: fix function def. order
...
epicsAtomicMemoryBarrierFallback used before definition
2015-03-09 19:48:30 -04:00
Andrew Johnson
e0f97a9215
Set testPlan in dbCaStatsTest.c
2015-03-09 18:07:13 -05:00
Andrew Johnson
a3793c368a
Merged changes from 3.14 branch up to revno 12561
2015-03-09 17:59:08 -05:00
Andrew Johnson
dad5b1cc94
Merged Michael's fixatomiclinkage branch
2015-03-09 17:43:44 -05:00
Andrew Johnson
20155602fc
Renamed the record test harness routine
...
Now named epicsRunRecordTests and as such it
matches the other test harness routines.
2015-03-05 14:50:42 -06:00
Andrew Johnson
f6067a3949
Merged changes from 3.14 branch up to revno 12557
2015-03-03 11:35:46 -06:00
Andrew Johnson
61438b4e7e
Fix Windows builds
2015-03-02 15:04:09 -06:00
Andrew Johnson
d432fb3115
Fix Darwin build
2015-03-02 14:22:25 -06:00
Andrew Johnson
02a105e5ab
Merged changes from 3.14 branch up to revno 12540
2015-03-02 11:43:20 -06:00
Andrew Johnson
e8c9783e06
Merged changes from 3.14 branch up to revno 12530
2015-03-02 10:35:35 -06:00
Andrew Johnson
fb1a0c2b1e
Clean up g++ warnings in repeater.cpp
2015-02-28 20:20:56 -06:00
Andrew Johnson
92b4d8e48d
Merged changes from 3.14 branch up to revno 12526
2015-02-27 18:11:37 -06:00
Andrew Johnson
b745dbff10
Merged changes from 3.14 branch up to revno 12516
2015-02-27 17:47:55 -06:00
Michael Davidsaver
89596fc308
Fix typo in epicsAtomicTest.cpp testClassify()
2015-02-27 17:29:11 -05:00
Michael Davidsaver
73dcc2745f
Fix for Solaris C++ compiler
2015-02-27 17:29:11 -05:00
Michael Davidsaver
332fd550ad
epicsAtomic: RTEMS ISR safe impl. for epicsAtomicLock()
2015-02-27 17:29:11 -05:00
Michael Davidsaver
214c9003a9
epicsAtomicTest: include some information about build conditions
...
Give some idea which implementations are being used
2015-02-27 17:29:11 -05:00
Michael Davidsaver
0d4519eb54
epicsAtomic: fix C linkage
...
With gcc -std=gnu99 (present default) functions qualified
with plain 'inline' emit strong public symbols (nm says 'T')
in every compilation unit the definition appears.
This causes linking to fail in all cases where executable
are statically linked, and many where dynamic linking is used.
2015-02-27 17:29:11 -05:00
Michael Davidsaver
84e9ff3bc5
libCom/test: Add epicsInlineTest
2015-02-27 17:29:10 -05:00
Michael Davidsaver
b3bfac3f16
add EPICS_ALWAYS_INLINE
2015-02-27 17:29:10 -05:00
Andrew Johnson
b7baa2504e
Only build Cap5.so for the host architecture
2015-02-24 16:23:14 -06:00
Andrew Johnson
4a0880003c
Support MinGW cross-builds on linux-x86 and x86_64
2015-02-24 14:30:42 -06:00
Andrew Johnson
2d7a8a4aee
Pull blockingSockTest.cpp from 3.15
2015-02-20 16:35:47 -06:00
Andrew Johnson
4cf3613ace
Applied missing-header.patch
2015-02-20 12:38:26 -06:00
Andrew Johnson
0fa8702842
Applied test-epicsTimerQueue-reference-counter.patch
2015-02-20 12:34:53 -06:00
Andrew Johnson
a437768c9c
Applied linux-runtime-disable-RT-priorities-if-not-supported.patch
2015-02-20 12:25:28 -06:00
Andrew Johnson
3e6597be90
Applied record-support-santity-check.patch
2015-02-20 12:13:05 -06:00
Michael Davidsaver
c0cf25eeee
dbCa: dbCaPutLinkCallback initialize entire array
...
CA links will write the full target array size.
Ensure that uninitialized elements are zeroed.
2015-02-19 19:18:27 -05:00
Michael Davidsaver
3501fda48d
dbCa: dbCaPutLinkCallback prevent out of bounds write
...
The internal buffer of the caLink is sized based on the
number of elements of the destination PV.
2015-02-19 19:18:16 -05:00
Andrew Johnson
e822d8d8c4
Suppress corrupt error output from dbStatic parser
...
The yyerror() routine gets called twice in some cases.
Don't print the yytext or input file location the second
time through -- yytext has already been freed, and the
user doesn't need to see the location twice.
Fixes lp:1422803
2015-02-19 17:33:27 -06:00
Andrew Johnson
9813fa6474
db: Allow link fields to return a DOUBLE value
...
A dbCa link does a ca_get with type DBR_CTRL_DOUBLE
to populate its list of attribute values immediately
after connecting. If the target is a DBF_*LINK field
it used to return an error, preventing the link from
properly connecting. This change makes dbGetField()
return a single NAN value instead of rejecting the
request.
Fixes: lp:545358
2015-02-19 15:55:48 -06:00
Andrew Johnson
804aeafaa7
Fix dbVerify()'s postfix buffer size
2015-02-18 16:44:07 -06:00
