GitLab

Fixed project_manifest.py to map dependency to tcmalloc (still used in
CMT config) to the new name of the package (gperftools).
This is a temporary hack needed to generate the RPM dependencies
correctly.

See merge request !180

`GaudiExamples/options/AsyncIncidents.py` demonstrates example utilization of asynchronous incident mechanism.

Fixes GAUDI-1213

See merge request !181

Fixed project_manifestproject_manifest.py to map dependency to tcmalloc (still used in CMT config) to the new name of the package (gperftools)

Fixes GAUDI-1210

See merge request !179

addresses the error reported in GAUDI-1210

if a DataHandle has a blank key (which in any case is not
allowed in the event store), ignore it when assembling the
data dependencies of the Algorithm or AlgTool

See merge request !169 and atlas/Gaudi!25

Fixes GAUDI-1202

See merge request !148

install Python files in cmake directories, because they
might be needed by CMake modules

See merge request !171

GCCXML is still mentioned in EnableROOT5.cmake because it's needed, but Gaudi,
anyway, cannot build with ROOT 5.

added a static function to EventIDRange which takes a
variable number of EventIDRange arguments, and returns the
intersection of them, but getting the maximum of the
`start()` and minimum of the `stop()`.

See merge request see atlas/Gaudi!22, !167

see atlas/Gaudi!21

This patch fixes:
 - setting fail flag in the EventContext for a stall
 - setting fail flag in the EventContext if an Algorithm
   throws an exception
 - don't unset the fail flag in the EventContext when an
   Algorithm hasn't failed

See merge request !165

Using std::list instead of std::vector allows the
possibility of listeners being added during the handling of
an incident, which cannot be forbidden because a handler may
trigger instantiation of a component that wants to be
registered.

From Scott Snyder. See atlas/Gaudi!20

See merge request !164

(from Attila)
The current implementation in these functions instantiates
duplicate tools when we declare a public ToolHandleArray as
a property to an algorithm or a tool. This modification
fixes the duplication in the ATN tests that the error
originally showed up in.

See merge request !158

See merge request !172

check inheritance from `TH1` instead of `TH11`

Fixes GAUDI-1209

See merge request !170 and atlas/Gaudi!26

install Python files in cmake directories, because they might be needed by CMake modules

The options --sh, --csh and --py report only the
differences, while the behaviour without a command is to
print all the environment.

The option --all can be used to get the whole list also with
the --<shell> options.

See LBCORE-1111, LBCORE-1112, lhcb-core/LbScripts!61

See merge request !163

Allow building of Gaudi with local externals.

It also contains a fix to an issue introduced with merge
request !160 (GAUDI-1212).

Fixes GAUDI-1211, GAUDI-1212.

See merge request !166

When dlopen files in listcomponents.exe, listcomponents
prints only that the library didn't load, but doesn't
include information on the underlying error. This change
adds the result of dlerror to the error message.

See merge request !159

Propagate failure of event to the EventContext for more cases, don't overwrite the failure of another Alg

(from Steve Farrell) see atlas/Gaudi!17

This merge fixes a serious undefined behavior condition in ThreadPoolSvc introduced in the thread-tool-terminate merge. The boost::barrier was moved to a function local which could drop out of scope before workers returned from waiting. Now, the barrier is again a class member of the ThreadPoolSvc. Tests are successful.
Additionally, I added a safety check in ThreadPoolSvc::terminateTool that returns FAILURE if the pool wasn't actually initialized.

Finally, I put the ThreadPoolSvc's raw pointer class members into unique_ptrs for automatic deletion.

See merge request !160

IncidentSvc: make ListenerList a vector instead of list to avoid issues when adding to it when iterating over it