9cb03c7015
Fix installation triggered already by build target
2011-01-30 19:43:51 +01:00
daef6c7676
SCons: remove unnecessary indirections $SRCDIR, $MODULES, $ICONDIR etc
...
these don't create real flexibility and make definitions less readable
2011-01-30 19:20:02 +01:00
abf1bc776b
SCons: remove all explicit target and install specifications
...
now superfluous, because our custom builder handles that automatically
2011-01-30 18:56:51 +01:00
609873d90b
switch to use the new (better) builder implementation
...
especially this means to use the common well-known names again,
like "Program" "SharedLibrary". The customisation now happens
invisible in LumieraEnvironment.
2011-01-30 17:00:15 +01:00
ab481a80b8
better implementation of Lumiera custom SCons builders, wrapping the predefined builders
2011-01-30 16:47:03 +01:00
014c22b40a
SCons: rework build directory configuration. All customisations to LuimieraEnvironment
2011-01-30 15:27:21 +01:00
9e56434c7e
SCons: start concentrating all custom builders into LumieraEnvironment
2011-01-29 23:09:02 +01:00
bc22ec7faa
Install: first preliminary working installation setup
...
the installed lumiera exe can even be started...
...well with a bit of cheating: you need to cd into the lib/lumiera
because the PLUGINPATH problem isn't solved yet
2011-01-29 16:45:22 +01:00
7993759f8e
SCons: remove the homebrew tarballer
...
that was a nice SCons learning exercise, but never used
moreover, packaing is not the concern of a build system
2011-01-29 14:16:15 +01:00
a6810957b4
Install: change organisation of target and output directory variables
...
now supporting a PREFIX in addition to the INSTALLDIR.
The latter is intended for package building
2011-01-29 14:10:19 +01:00
2bcc8d9ae3
SCons: build loadable modules immediately to the target destination
2011-01-29 02:33:13 +01:00
35953b335b
SCons: rearrange output directory to target/modules
2011-01-29 02:06:21 +01:00
764a38abe6
SCons: experimental support for some library lookup concerns
...
- setting -rpath with $ORIGIN to build a relocatable package
- fix missing DT_SONAME (likely just a problem of very old SCons version)
2011-01-28 23:31:00 +01:00
a29591c299
SCons: several small improvements, e.g. valgrind-suppressionfile
2010-07-26 03:24:15 +02:00
002f024ed8
SCons: improve dependency handling for the testsuite
2010-06-26 05:08:16 +02:00
7fbf84dd5d
Bump required nobug version to 201006.1
2010-06-03 04:41:51 +02:00
f0aed2c2cc
fix dummy compilation/tests
2010-05-21 03:50:41 +02:00
86bc30a10a
Merge Ubuntu/Lucid adjustments. Drop pre 1.0 compatibility
2010-05-21 03:43:09 +02:00
0977a7de6d
SCons: replace deprecated Options() by new Variables() object
...
dop-in replacement, no functionality change. Requires SCons >= 1.0
2010-05-21 01:59:37 +02:00
d10380d124
SCons: dropping pre 1.0 compatibility
...
note: current SCons development is targetting 2.0,
current stable is 1.3, which is included in Squeeze
even Debian/Lenny includes 1.0.
2010-05-21 01:59:26 +02:00
3ca17af42f
SCons: new NoBug version 201005.1 and linking against -lrt for clock_gettime
...
Fixes #618 and #619 for SCons build
2010-05-15 02:20:15 +02:00
Odin Hørthe Omdal
792a595041
Fix for old GAVL versions (missing uint64_t)
2010-05-15 02:20:07 +02:00
Christian Thaeter
9d99300841
FIX : #619 , New Nobug required! version 201005.1
...
add the 'extra' argument to mpool dumps.
Solves Linking problems on some distros.
2010-05-12 00:46:39 +02:00
Christian Thaeter
5792d330b3
FIX #618 : test if -lrt is required for linking
...
This checks for clock_gettime availability in the clib
and in librt and adds -lrt to LUMIERA_LIBS when the
function is available in librt.
The check is only done for autotools based builds,
scons may need something else.
Other functions from the realtime need to be added on
demand.
2010-05-11 18:41:19 +02:00
030320352d
Bugfix test.sh: detecting literally defined output
2010-04-16 02:23:25 +02:00
9f2affb41b
SCons: feed test.conf to test.sh (for correct output filtering)
2010-02-15 03:25:01 +01:00
f75964c5ba
NoBug February-Release 201002.1 now required
2010-02-15 00:50:16 +01:00
5b9aa5deb2
landing the new threadpool implementation
2010-02-15 00:48:52 +01:00
1c63a02e23
augment and round up the C++ thread wrapper
2010-02-14 23:39:15 +01:00
90d311dc1d
Error integration: error::Flag as subclass of State. Wrapped throw
2010-02-13 20:56:41 +01:00
acd0671390
clarify some comments regarding GUI start
2010-02-13 17:41:16 +01:00
4e31db5a74
my private convenience wrapper for running an exe with valgrind
2010-02-13 06:01:49 +01:00
dc991ca563
valgrind suppression: add some more cases to be filtered
2010-02-13 06:00:38 +01:00
b41bd20de4
push accessing the DiagnosticContext down one layer
...
accessing the DiagnosticContext now inline when
providing the paramters for calling the C-functions.
No change in functionality, but saves us a lot of
syntactic noise.
2010-02-13 05:33:08 +01:00
3466793976
repackage as a more general facility (DiagnosticContext)
...
implementation unaltered (just managing the NoBug handle)
2010-02-13 04:54:59 +01:00
4dfd7266b9
draft solution to factor out management of the resource handle
...
based on the idea of a diagnostic context stack
2010-02-13 04:03:27 +01:00
Michael Ploujnikov
433448ff44
use _join() instead of a second _sync() to wait for the worker to finish
2010-02-12 07:20:32 -05:00
763f86fe0e
back out changes I do not want in lib/sync.hpp
...
especially, I do not want to pass a resource handle
through all locking function APIs; the memory
management of the resource tracker should better
be kept separate and not mixed with the monitor.
Also, I am rather reluctand regarding any extended
functionality for the monitor, like timed locks
or trylocks or read/write monitors. I think, the
monitor pattern is only beneficial when it is kept
fairly simple, advanced thread programming should
be pushed out into lib functions in the backend.
2010-02-12 05:22:17 +01:00
b556343fb6
re-enable typed-counter-test; passing just fine again
2010-02-12 01:54:12 +01:00
52a7d6993d
SCons: remove GThreads from linking (except for the GUI)
2010-02-11 03:19:42 +01:00
dfd70c6069
replace GThreads by the Lumiera thread wrapper
2010-02-11 03:06:42 +01:00
Michael Ploujnikov
d80ae17ea4
test simple-sync: avoid a race condition in output checking by using an assert
2010-02-10 20:53:44 -05:00
Michael Ploujnikov
62b70bfd7e
test sync-joinable: use asserts instead of checking the program output
2010-02-10 20:53:32 -05:00
Michael Ploujnikov
5d929f1522
test sync-many: replace output checking with asserts
...
Also modify the input value right after passing to a worker thread. This is a better test of synchronization.
2010-02-10 20:53:13 -05:00
Michael Ploujnikov
5606eb2d83
Merge remote branch 'ct/backend_devel' into second-tp-attempt
...
Conflicts:
src/lib/error.c
2010-02-09 07:29:43 -05:00
Michael Ploujnikov
25ec9a91f7
don't forget to free the error struct
2010-02-09 07:21:58 -05:00
Christian Thaeter
26f88bc3f1
Merge remote branch 'plouj/second-tp-attempt' into backend_devel
...
* plouj/second-tp-attempt:
basic joinable thread test
joinable thread sync test
2010-02-08 18:20:05 +01:00
Christian Thaeter
80e2db4800
hook the resourcecollector into safeclib on backend startup
2010-02-08 17:47:26 +01:00
Christian Thaeter
d103346482
hook mpool into the resourcecollector and using safeclib allocations
2010-02-08 14:28:54 +01:00
Christian Thaeter
9feb01e4fe
add hooks for malloc/free to mpool, add (no-op) purge function
...
We will need this to hook the resourcecollector in
2010-02-07 20:53:51 +01:00
