A bug in the delete function in hash.c was preventing the unit test of the new set ID API functions from running properly. In fixing this bug the entire hash table code was refactored to make it look more like the mempool service routines. Also the need to copy the string passed into the table's insert function was eliminated.
Adds new options HEADERROR and FLOWCHANGE to provide more rigorous criteria for hydraulic convergence. Also breaks HYDRAUL.C into 3 separate files to improve code readability.
this quite sizable commit does several things, but is primarily
focussed on building a toolkit that can run simultaneous
simulations/analyses within a shared memory space. Versions <=2.1 use a
long list of global variables that prevent multiple instantiations on
linux systems without resorting to compilation tricks (like duplicate
binaries or similar via static linkage). This version uses a single "Project" pointer to encapsulate
the network and analysis data. There are no changes to existing algo
implementations other than to accomodate dereferencing of the passed-in
pointers. A more detailed list of major changes below:
- mirrors all “ENxxxx” function calls with “EN_xxxx” versions (note the
underscore) that take an extra first parameter: a pointer to an
EN_Project struct, which contains all network, hydraulic, quality, and
associated data.
- tweaks some code formatting to make it more readable
- removes some deprecated/commented code that was sufficiently old
- fixes implicit type-cast warnings
* Added ENaddnode and ENaddlink functions
* More memory reallocations
* Added ENInit, ENsetheadcurveindex
* Added ENdeletelink and ENdeletenode
* restored default behavior for float types
* fixed type
* Added docstrings for ENinit
* cleanup change
* moves global rule variables to vars.h
* migrates rule structs to typedefs
for better readability
* char types to proper enums
fixes#93
* Change some variable declarations for compatibility
Changes to keep compatibility with C89 compilers: variables must be
declared at the top of the functions. Remove the use of EN_LinkType in
function call as it is not compatible with ENgetnodetype.
* Moved declaration of idstodelete to top of function
* Updated ENinit function and headers
Updated header files with new functions
Updated def file with new functions
For ENinit changes names of parameters #98
Added enum for headloss formula
* Missed these files in 1a033fc
* migrates char types to enums fixes#93,
supports unified link/node type enums, rather than public/private
redefinitions
* removing links in reverse-index order maintains proper indexing fixes#96
* style
* clarifies curve getter units issue (dox)
closes#95
* fixes link/node confusion in ENsetlinktype
partially reverts a3bce95dc330a5a297634a303d438e2e1bc41cc9
* partial compilation fix
* fixes dox issue
* fixes allocation issues with enums
- updates style in various places
- introduces FlowDirection enum
- use snprintf to prevent overflow
* fixes enum type cast
* updated mac project settings
* Use of _snprintf on Windows and remove DLLEXPORT from mempool.h
snprintf it not compatible on Windows so we use _snprintf
mempool gave starnge compilation errors while removing DLLEXPORT worked.
Not sure why these functions needed to be exposed in the DLL?
* Revert "Use of _snprintf on Windows and remove DLLEXPORT from mempool.h"
This reverts commit 6238f77d47fa0feaabe5836043c006937de433a2.
* use of _snprintf instead of snprintf on Windows and removed DLLEXPORT from mempool.h
Had compilation errors on mempool.h. Removed DLLEXPORT so solve it. Not
sure why there was a need to expose these functions?
* Shift indices for Links in ENaddnode
Need to shift indices for Links not just Pipes since a pump could be
connected directly to a reservoir. Also set the defult base demand to
zero (was 5).
* Set defualts for madatory link properties in ENaddlink and small fix for ENsetheadcurveindex
Relates to #102 and closes#103
* wraps globals into structs, duplicates api functions with objective versions
* parse and serialize Comment field for network elements
related to #47
* adds getter for head/efficiency curve
in EN_getlinkvalue
* adds getter for event node index
… to return the index of the junction (tank) that triggered the event.
* fixes edge case in parsing
… where inp files without demands in [JUNCTIONS] and without any
[DEMAND] categories will fail.
* adds freeing function for project pointer
* removes redundant string literals, fixes overrun issue in error message getter function
* check for hydraulics already closed
* moving error definitions to data file
* deprecates ENR err message getter (unused)
* updates location of errors data file
also begins to expose blind structs to curves and patterns,
anticipating buildout of APIs for those.
* updates CLI output to reflect executable name as invoked
relates to #109
* Feature nrtest (#131)
* Initial commit EPANET testing tools.
* Initial commit for epanet-nrtestsuite
* SWIG wrapper for EPANET outputapi (#118)
* Removed pervious version of outputapi and wrapper
* SWIG wrapper for EPANET outputapi
* Patching cmake build script fixed target for outputapi
* Build failing on deprecated test script
* Minor changes. Responding to review comments.
* Feature nrtest (#121)
* Configured python setup to automatically build nrtest tools.
* Working on build / testing automation
* Adding EPANET 2.0.12 benchmark
* Updated Travis yml to run nrtest
* Fixing InsecurePlatformWarning
* Fixing InsecurePlatformWarning again
* Fixing InsecurePlatformWarning
* Fixing InsecurePlatformWarning
* Fixing InsecurePlatformWarning
* Update .travis.yml
* Update .travis.yml
* Update .travis.yml
* Update .travis.yml
* Working on configuring python environment and building test tools under Travis CI.
* Making gen-config.sh and run-nrtest.sh executable
* Debugging .travis.yml
* Debugging .travis.yml
* Debugging .travis.yml again
* Debugging .travis.yml again
* debugging travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* debugging Travis setup
* Fixing bug with __strncpy_chk destlen < len
* nrtesting clean up
* re-implements fixes from:
5eead5ae403c788567a4
* removes extraneous build files, moves cmake and updates travis
* mirror of 9b37035560f9683f1514b439f7586a5c17bca5bf
* Move some variable declarations
* More variable declarations
* Fix TmpDir
* Allocate _defaultModel
* Fix EN_addcurve funcrion
* Fix for inpfile
* Fix writeRuleinInp call
* Set MAXMSG to 79 chars
* Fix for flow direction
* Refactoring testing related python packages and SWIG wrapper bug fix (#139)
* Eliminated epanet-reader package. Removed numpy dependency from epanet-output. Fixed reference counting bug in SWIG wrapper. Added error checking to run_nrtest.sh. Added nrtest package to requirements file.
* changing buildhome directory
* Fixing bug related to preprocessor definition of PI
The code has the following modifications:
- RULE section is written just after controls
- SYST is now written SYSTEM (also PRESSURE, DEMAND, POWER, SETTING,
FILLTIME and DRAINTIME have been changed)
- some description to the function writeRuleinInp has been added
- Actions are now written in the correct order
Added the new funcions to the DEF file.
Updated VB headers from Double to Single
Changed a few APIs calls to support EN_API_FLOAT_TYPE - this needs more
work for the rest of the new functions
this version works, at least with Net3, simplenet, and SampleTown -- but the results are non-canonical, as the current workaround for the issue above writes tiny flow values to the binary file for closed elements. presumably we should check for closed status in the transport() function.
