Neil Brown
cdcbebd2ae
Fixed the instances generated for Maps
2008-12-14 15:38:01 +00:00
Neil Brown
263dfd6fd3
Allowed the different instances for the Polyplate classes to be written out to separate files (to help GHC's memory usage down a bit)
2008-12-14 15:37:05 +00:00
Neil Brown
cd9b8462b4
Made Tock depend again on the BUILT_SOURCES (which I broke in an earlier patch)
2008-12-14 15:36:30 +00:00
Neil Brown
d24a984442
Fiddled with the PolyplateSpine class and added code to generate the instances
2008-12-09 16:18:06 +00:00
Neil Brown
189032e275
Fixed up the full traversal of the tree to allow specification of a default value
2008-12-09 11:20:39 +00:00
Neil Brown
a7f04a030c
Added more stuff relating to the spine view in Polyplate, but now reconsidering its necessity
2008-12-05 16:48:16 +00:00
Neil Brown
bcbd18e5b7
Changed the build system, since GenNavAST now (indirectly, via CompState) relies on OrdAST existing
2008-12-03 17:39:04 +00:00
Neil Brown
24a7b7aba8
Changed GenNavAST to also generate instances for CompState
2008-12-03 17:37:29 +00:00
Neil Brown
c83c383181
Fixed the GenClassPerType option to actually generate the forwarding instance from PolyplateM to the type-specific class
2008-12-03 17:36:40 +00:00
Neil Brown
360b076be9
Made genMapInstance and genSetInstance generate instances for the directly contained types
2008-12-03 17:30:57 +00:00
Neil Brown
29b79000f1
Made the instance generation obey the settings for GenClassOption
2008-12-03 17:19:16 +00:00
Neil Brown
24364c0cbe
Pulled out the final bits of the instance generation that I want to make configurable
2008-12-03 17:07:43 +00:00
Neil Brown
7ee2475363
Pulled out more commonality, this time relating to contexts
2008-12-03 16:55:48 +00:00
Neil Brown
14d6dec1d1
Pulled out some of the details of generating a Polyplate instance into a helper function, to help make it easier to configure
2008-12-03 16:41:01 +00:00
Neil Brown
624e10b9ff
Added a special case for sets, and some documentation explaining why this mechanism is needed
2008-12-03 15:42:52 +00:00
Neil Brown
85f917df2f
Added a mechanism for having special cases for awkward types (such as Map) during the instance generation
2008-12-03 15:32:43 +00:00
Neil Brown
890e7ea9a6
Fixed many of the modules to use the new Polyplate-based Traversal system
2008-12-02 18:06:10 +00:00
Neil Brown
a3bcb32937
Switched the Pass and Traversal modules over to Polyplate, breaking most of the other modules
2008-12-02 17:33:55 +00:00
Neil Brown
e78a30c231
Added an applyBottomUp function (non-monadic)
2008-12-02 17:17:22 +00:00
Neil Brown
6a0529a9ce
Took out the pregen sources from tock_SOURCES_hs, which I think would have caused them to be packaged (don't think we would want to, or at least we don't need to)
2008-12-02 17:06:39 +00:00
Neil Brown
4b9d186a65
Added the parameter to increase the context stack (for working with polyplate)
2008-12-02 17:06:17 +00:00
Neil Brown
44026e985a
Adjusted the clean rule to fit with the new Polyplate modules
2008-12-02 17:06:00 +00:00
Neil Brown
4fa324cbfe
Changed polyplate so that users should always supply their own headers (since it must have any appropriate imports)
2008-12-02 16:04:04 +00:00
Neil Brown
69e5d1d20d
Reworked GenInstances, and wired in the option relating to overlapping instances
2008-12-02 15:50:38 +00:00
Neil Brown
c9100fb467
Added the first version of GenInstances (taken from my latest improvements) and switched GenNavAST to use it, but it needs tidying up and to take account of the options it is given
2008-12-02 15:16:22 +00:00
Neil Brown
01fff6e617
Realised listify should return a list, and added depth-first and breadth-first versions
2008-12-02 14:18:11 +00:00
Neil Brown
b6176ee615
Added functions and schemes related to the new query functions
2008-12-02 14:10:43 +00:00
Neil Brown
a2063da839
Added some thoughts (and a new type-class) concerning possible uses of Polyplate
2008-12-02 13:49:36 +00:00
Neil Brown
82cae7691a
Fixed the remaining missing functions to get the Schemes compiling
2008-12-02 13:13:54 +00:00
Neil Brown
77251b842c
Added the Polyplate modules to the Tock build system and fixed various errors
2008-12-02 13:08:38 +00:00
Neil Brown
8b28fee077
Started a module (Schemes) containing various Polyplate traversal schemes
2008-12-02 12:59:17 +00:00
Neil Brown
df2b85ea27
Moved Polyplate into a sensible package (based on the Uniplate module names)
2008-12-02 12:26:12 +00:00
Neil Brown
e7231d8959
Added the first version of the standalone Polyplate module
2008-12-02 12:24:17 +00:00
Neil Brown
9403f4d68b
Filled in more expressions and corrected the generation of byte literals by borrowing code from the C backend
2008-11-27 09:25:34 +00:00
Neil Brown
ce86ec2337
Changed genMissing so that it generates a run-time error in the generated code
2008-11-27 09:25:07 +00:00
Neil Brown
e79ccf1b46
Fixed it so that names beginning with an upper-case letter are prefixed by an underscore in the CHP backend
2008-11-27 09:24:31 +00:00
Neil Brown
769ed4f3ca
Filled in a little more of the CHP backend
2008-11-26 20:39:51 +00:00
Neil Brown
70d7b941cc
Added support for more types and some expressions (mainly array stuff)
2008-11-26 18:47:32 +00:00
Neil Brown
ab8fd2434c
Made all the missing stuff call genMissing
2008-11-26 15:51:21 +00:00
Neil Brown
d962126fed
Added the framework for variable declarations
2008-11-26 15:47:22 +00:00
Neil Brown
d7c61a70c0
Fixed some indentation issues and added markers for things that are missing
2008-11-26 15:37:49 +00:00
Neil Brown
5d0391c98d
Fixed it so that processes at the top-level are not inside let bindings
2008-11-26 15:28:36 +00:00
Neil Brown
caff6ae702
Added some more types, and import declarations
2008-11-26 15:27:16 +00:00
Neil Brown
b3c1a6f13b
Filled out the very basics of a CHP backend, enough to get some proc headers
2008-11-26 15:11:51 +00:00
Neil Brown
e3c426e870
Added full-compile support for the CHP backend
2008-11-26 12:48:48 +00:00
Neil Brown
78716a2727
Added a CHP backend option to Tock
2008-11-26 12:24:21 +00:00
Neil Brown
aa10b0113d
Allowed direction specifiers to be added to variables that are channel bundles when they are passed as arguments
2009-04-12 13:51:48 +00:00
Neil Brown
da43dcb3f5
Allowed multiple empty dimensions in mobile types, not just one
2009-04-12 13:51:32 +00:00
Neil Brown
5534d1c3d6
Fixed some problems with DATA TYPEs that were actually MOBILE stuff not being resolved at the right points
2009-04-12 13:50:59 +00:00
Neil Brown
47d565a3b9
Changed records to always use their original name in the C code
...
This fixes the problem with PROCs complaining that one munged version of the record name was not the same as another.
This will cause problems if two different record definitions for the same name are visible in the same file.
2009-04-12 13:06:49 +00:00
