summaryrefslogtreecommitdiffstatshomepage
path: root/src/emu/netlist/plib
Commit message (Collapse)AuthorAgeFilesLines
* Moved netlist from emu to lib (nw) Miodrag Milanovic2015-09-1213-4388/+0
|
* Some pedantic netlist changes. (nw) couriersud2015-08-282-11/+124
|
* Cleanups and version bumpmame0165 Miodrag Milanovic2015-08-264-10/+8
|
* Moved nltool.c and nlwav.c into src/emu/netlist/prg. couriersud2015-08-237-53/+31
| | | | Added src/emu/netlist/build/makefile. This allows netlist to be ripped out of the tree and to compile it standalone. (nw)
* Created a separate logging class. Netlist code should now be at least couriersud2015-08-236-56/+258
| | | | | 98% type safe. No more fuzzing around with SIZEFMT and friends. Changed formatting to use python style format strings. (nw)
* Fix netlist build (nw) Vas Crabb2015-08-171-2/+2
|
* Added seconds() and attoseconds() to attotime and prefixed members with couriersud2015-08-151-1/+1
| | | | | | | m_. Rewrote code accessing members to use seconds() and attoseconds(). The changes were triggered by a test how gcc __int128_t would perform as the internal representation. This test revealed that the current implementation is still faster. (nw)
* Some netlist_time maintenance. Straightened code using it and added couriersud2015-08-143-0/+22
| | | | | support for 128 bit resolution on systems supporting it. This is however disabled since all platforms I know don't have native support but emulate 128 bit. (nw)
* Changed UINT16 .. UINT64 to generic types in pformat. couriersud2015-08-132-31/+133
| | | | | | @MSVC Users: I checked the MSVC++ documentation prior to committing this. I also checked on a MSVC2010 installation that "long long" and "%lld" actually work at least in MSVC2010. Please let me know if this now works. (nw)
* Make more use of pformat. (nw) couriersud2015-08-133-23/+18
|
* Fix emscripten compile (nw) Justin Kerk2015-08-121-1/+1
|
* Remove vsscanf from netlist. Change default format for double to "g". couriersud2015-08-112-13/+2
| | | (nw)
* Fix mingw compiles (both 32 and 64). These seem to have a different couriersud2015-08-111-2/+3
| | | understanding of size_t than *nix compilers. (nw)
* Temporarily fix compile by commenting line 395 and kepping 409 in file ↵ yz70s2015-08-111-1/+1
| | | | src/emu/netlist/plib/pstring.h (nw)
* Fix clang issues. (nw) couriersud2015-08-102-5/+5
|
* Make netlist more typesafe. Added a pformat class to reduce sprintf couriersud2015-08-103-2/+85
| | | | usage. The approach is also suitable for translated strings with arbitrary positioning of parameters. (nw)
* utf8 support for pstring. Opted for a scalable solution which should be couriersud2015-08-107-250/+527
| | | | | easily extensible to utf16 and utf32 as well. All position related operations now operate on char code positions instead of byte positions. [Couriersud]
* More usage of streams and aligned exception use. (nw) couriersud2015-08-107-12/+37
|
* Added simple stream classes to netlist code. (nw) couriersud2015-08-036-106/+698
|
* Cleanups and version bumpmame0164 Miodrag Milanovic2015-07-294-17/+15
|
* Consolidate the scattered model parsing code. Now models support couriersud2015-07-232-1/+4
| | | | recursive models,e.g. DIODE("1N914(IS=1e-15)"). Removed ".model" from model definitions. (nw)
* please bear with me.. these are needed to compile in GCC 5.2 and I want ↵ Cowering2015-07-232-0/+24
| | | | mingw guys to see them. with luck they can be reverted in a couple of days (nw)
* More phashmap_t usage. Quite some nice effect on larger netlist startup couriersud2015-07-201-20/+35
| | | | time. Hash objects can now deal with arbitrary hash width. (nw)
* Use hashmap for factory searches. (nw) couriersud2015-07-201-11/+0
|
* Changed phashmap_t to dynamically resize hash table. (nw) couriersud2015-07-191-5/+86
|
* Aligned builtin device identification with parser device couriersud2015-07-191-2/+0
| | | identification. (nw)
* Added a phashmap_t object for faster lookups. (nw) couriersud2015-07-192-0/+141
|
* Add macro library for opamp models. (nw) couriersud2015-07-161-0/+8
|
* Fix issues with gcc4.0 mentioned by Barry. (nw) couriersud2015-07-091-1/+1
|
* Identify size_t usage in printf by enclosing the argument in couriersud2015-07-022-1/+11
| | | | SIZET_PRINTF(). This is typesafe, can be identified and allows to use proper 64bit computing with compilers up to it. [Couriersud]
* Netlist now compiles with couriersud2015-06-307-66/+95
| | | | | | | -std=c++98 -Wall -Wpedantic -Wextra -Wno-long-long -Wno-variadic-macros and -std=c++11 -Wall -Wpedantic -Wextra [Couriersud]
* Hopefully fix more "maybes" which maybe popup sometime in the future. couriersud2015-06-282-3/+8
| | | Maybe. (nw)
* Code maintenance. In addition, added a parameter to select iterative couriersud2015-06-241-1/+1
| | | solver. (nw)
* Cleanups and version bumpmame0163 Miodrag Milanovic2015-06-242-11/+10
|
* Some more moves to the netlist namespace. (nw) couriersud2015-06-192-1/+14
|
* From 45% to 60% to 99%. That's the improvement achieved for a 89x89 couriersud2015-06-101-0/+18
| | | | audio matrix mostly solved by elimination. Cleaned up some code as well. [Couriersud]
* Another go at fixing gcc 4.4 couriersud2015-06-091-0/+1
|
* Made it more convenient to add frontiers (impedance barriers). These couriersud2015-06-081-1/+11
| | | | | | - at the right place - enable netlist to split a netlist into multiple smaller netlists. Side mark: The "old" discrete system assumed such a device in every node. (nw)
* Fixed a number of bugs and some additions: couriersud2015-06-072-0/+11
| | | | | | | | - fixed bugs in the spice(Kicad) conversion - fixes submodel difference between inline and parsed netlist - added truthtable 7404 and 7486 models - aligned input and output naming for truthtable and specialised 74xx models.
* Found some cycles by removing static declarations in some functions. couriersud2015-06-031-3/+3
| | | (nw)
* Converted netlist into a library. Moved most code from nl_convert.h to couriersud2015-06-011-2/+6
| | | | | | | nl_convert.c. Targets using netlist must now specify this explicitly with MACHINES["NETLIST"] = true Added subtarget "nl" which only contains games with netlist elements. (nw)
* Code maintenance. Remove commented out directives. (nw) couriersud2015-05-311-1/+1
|
* Netlist now uses the same approach as delegate.h to derive member couriersud2015-05-316-14/+116
| | | | | function pointers. If the platform doesn't support this approach, virtual function calls will be used. In addition, this commit contains modifications for standalone compile. (nw)
* Fix OpenMP by using spinlocks in push_to_queue. However no observable couriersud2015-05-315-81/+82
| | | | performance improvement. Without OpenMP, locks will not be compiled. Removed a couple of inlines and added consts were appropriate. (nw)
* Commit omitted nl_convert.h. Renamed "find" to find_by_name in couriersud2015-05-291-3/+3
| | | pnamed_list for clarity. (nw)
* First version of a eagle to netlist converter. (nw) couriersud2015-05-292-2/+27
|
* Got truthtable parsing working in the parser. The plib preprocessor now couriersud2015-05-282-4/+58
| | | | | defines __PLIB_PREPROCESSOR__. This can be checked in netlist code and allows even more sharing code between compiled (embedded) netlists and external netlists. (nw)
* Introduced user definable truth tables: couriersud2015-05-271-4/+7
| | | | | | | | | | | | | | NETLIST_START(lib) TRUTHTABLE_START(TTL_7400A_NAND, 2, 1, 0, "+A,B") TT_HEAD(" A , B | Q ") TT_LINE(" 0 , X | 1 |22") TT_LINE(" X , 0 | 1 |22") TT_LINE(" 1 , 1 | 0 |15") TRUTHTABLE_END() NETLIST_END() This enables the addition of devices without changing the netlist source code and allows the creation of libraries. Used pong.c as a proof of concept for the time being. [Couriersud]
* Moved all files in src/emu/netlist starting with p into plib folder. couriersud2015-05-2711-0/+2469
This is a first step to ease synchronisation with a stand alone, e.g. outside mame, netlist implementation. More signed/unsigned cleanups and started work on generic truthtable devices. (nw)