[Yhc] Changes to build/SConscript.libffi, prepend include path
Goetz Isenmann
G.Isenmann at science-computing.de
Fri Nov 10 04:44:15 EST 2006
Hi!
I got errors during the ctypes libffi build, because the preprocessor
found ffi header files in /usr/local/include before those shipped with
ctypes.
Prepending the path to the internal header files instead of appending
it fixed this issue for me.
--
Goetz Isenmann
-------------- next part --------------
New patches:
[look for internal libffi headers first
Goetz Isenmann**20061110092325] {
hunk ./build/SConscript.libffi 11
- ffienv.Append(CCFLAGS="-DX86_WIN32")
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi_msvc"])
+ ffienv.Prepend(CCFLAGS="-DX86_WIN32")
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi_msvc"])
hunk ./build/SConscript.libffi 15
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
hunk ./build/SConscript.libffi 18
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
hunk ./build/SConscript.libffi 21
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src/x86", "#depends/ctypes/libffi"])
hunk ./build/SConscript.libffi 24
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src", "#depends/ctypes/libffi/src/powerpc", "#depends/ctypes/libffi"])
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src", "#depends/ctypes/libffi/src/powerpc", "#depends/ctypes/libffi"])
hunk ./build/SConscript.libffi 27
- ffienv.Append(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src", "#depends/ctypes/libffi/src/powerpc", "#depends/ctypes/libffi"])
+ ffienv.Prepend(CPPPATH=["#depends/ctypes/libffi/include", "#depends/ctypes/libffi/src", "#depends/ctypes/libffi/src/powerpc", "#depends/ctypes/libffi"])
}
Context:
[Attempt to fix the extra-spaces in ghc command line once again.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061108150229]
[Import System.FilePath.Version_0_09 instead of just System.FilePath, this gives better assurances of reverse compatability
Neil Mitchell**20061108143853]
[Fix the FilePath build script to build the Version files as well, this is now specific to each minor revision of FilePath, so probably needs fixing properly
Neil Mitchell**20061108143819]
[Detect if the user is using an rc release of darcs.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061108141640]
[Update the Echo demo program as Runtime has been updated
golubovsky at gmail.com**20061108034658]
[More primitives in Overlay, worked around oversaturation
golubovsky at gmail.com**20061108033621]
[Fix sExp to not crash, this fixes bug 75
Neil Mitchell**20061107141604]
[Fixed the Defect #72 (oversaturated function calls)
golubovsky at gmail.com**20061107045634]
[Move more towards Control.Monad.State
Neil Mitchell**20061107025938]
[Remove applyStateMonad from Util.Graph, was unused
Neil Mitchell**20061107025400]
[Use execState/evalState instead of runState in some places
Neil Mitchell**20061107025138]
[Use runState instead of applyStateMonad
Neil Mitchell**20061107024852]
[Remove getState and setState, get and put are used instead
Neil Mitchell**20061107024338]
[Remove the qualification from Control.Monad.State (offending name clashes have been fixed now)
Neil Mitchell**20061107024005]
[In the ByteCode.CompileLib rename lift to liftC
Neil Mitchell**20061107023854]
[Rename lift -> liftC in DotNet.Compile
Neil Mitchell**20061107023715]
[Rename State to CState in the compiler
Neil Mitchell**20061107023505]
[Import Control.Monad.State, use State instead of StateMonad for the type
Neil Mitchell**20061107023134]
[Rewrite the state monad in terms of Control.Monad.State
Neil Mitchell**20061107021843]
[Add -package mtl in the Haskell module - this is a hacky patch and needs doing properly by Andrew!
Neil Mitchell**20061107021500]
[Add type signatures for trimLeft and trimRight, the last missing two
Neil Mitchell**20061107020907]
[Add type signatures to Util.Text
Neil Mitchell**20061107015308]
[Make the StateMonad export less of its internals
Neil Mitchell**20061107012914]
[Fix the play instance when encountering a let, it was going too deep on the RHS of a let - missing top level elements
Neil Mitchell**20061105160335]
[Fix type. Remove unneed code from build system.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061105113633]
[Stole replicateM, replicateM_ from GHC Control.Monad
golubovsky at gmail.com**20061105034313]
[First demo added: Echo
golubovsky at gmail.com**20061104055934]
[Remove the -i options, forgot them when I removed -d
Neil Mitchell**20061103145201]
[Revert Dimitry's patch, by fixing the file path semantics in Flags (which means that included source files are found correctly)
Neil Mitchell**20061103142321]
[Do not specify the -d flags any more, Yhc uses paths as relative to the file you are compiling, not relative to the current directory
Neil Mitchell**20061103142044]
[Make sure the rootPath is represented in the location of a file, required to get include's working correct (may well break something else...)
Neil Mitchell**20061103123416]
[Get closer to building yhi with ghc on Windows.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061103092321]
[Add scons test quick command for testing things quickly.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061102164657]
[Allow ghc to be used to compile yhi. Pass cc=ghc to scons to activate.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061102162937]
[Remove warning about libraries not being recognised when it is.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061102155855]
[Allow using ghc as compiler for yhi.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061102155535]
[Allow scons build libraries. Fail on unrecongised options.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061102155228]
[Dirty hack to Core Linker - set path to ycr same as to hs
golubovsky at gmail.com**20061103043047]
[Remove all the bootstrap files, now scons can build them properly
Neil Mitchell**20061102151137]
[No need to copy Ratio.hbc, its made automatically by the build system
Neil Mitchell**20061102151004]
[Make sure that you also search the includes directories passed on the flags, fixes a bug reported by Dimitry
Neil Mitchell**20061102133630]
[Add main x y = Main.main x y, so that reachability knows where to start from
Neil Mitchell**20061102003151]
[Use not . Set.member, instead of Set.notMember - notMember is only in GHC 6.6. and Hugs Sep 2006 (so no one but me has it!)
Neil Mitchell**20061102002222]
[Add Yhc.Core.Overlay, for Catch and the JS back end
Neil Mitchell**20061101172331]
[Fixed the unmarshaller for characters so it works with computed chars
golubovsky at gmail.com**20061031123437]
[Simplify the fix to finding the version number.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061031121535]
[Check we can find matching .hs file for a required .o file.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061031121507]
[use_records
golubovsky at gmail.com**20061031033006
Changed the code to use record syntax for Core and CoreFunc.
]
[Move back towards the old let method, accepting that let's are really letrec's under the covers - this fixes repeat using a where binding
Neil Mitchell**20061031021937]
[Fix a bug in the let bindings, let's can be written as let {x = y; z = x} in q, where the z refers to the previous x - i.e. letrec's - make this more explicit
Neil Mitchell**20061031020608]
[Change the serialise code so checking for the right version is done sooner, and more likely to get an error here rather than in the Binary instance
Neil Mitchell**20061031011608]
[Allow the -corelink module not to be called module Main, and therefore allow alternative type signatures on main
Neil Mitchell**20061031005515]
[Only remake the .yca file if the source file has changed (not enough, should be on any recompilation, but hard to express for now)
Neil Mitchell**20061031004710]
[Base the location of the .yca file on the ycr file location, not on the source file
Neil Mitchell**20061031004325]
[Change the -hide flag so it gives cleaner semantics for Catch, allowing multiple Main files in one directory
Neil Mitchell**20061031003849]
[Ensure that the directory for the .ycr file exists
Neil Mitchell**20061031003713]
[Remove the makefile, was long since out of date
Neil Mitchell**20061030140939]
[unmarshaller_deep_eval
golubovsky at gmail.com**20061030115448
Added deep evaluation for arrays by applying `exprEval' to the head element
of each cons cell (was applied only to the tail element before).
As a result, nested lists of strings are unmarshalled correctly,
and toString called on [["aaa","bbb","ccc"],["dd","ee"],["f"]]
returns 'aaa,bbb,ccc,dd,ee,f' (Javascript itself does not place square
brackets when doing Array.toString).
]
[string_array_unmarshaller
golubovsky at gmail.com**20061030050107
Implemented unmarshaller (i. e. Haskell -> Javascript) for strings
and lists (represented as Javascript arrays). Works for lists of integers
and strings (literals and computed), but for e. g. ["aaa","bbb","ccc"]
Javascript sees an array of three objects (deep expression evaluation is needed).
Core to Javascript converter may have lost its compatibility with
the latest core types change introduced together with the core linker.
]
[Add type signatures for Core.Linker
Neil Mitchell**20061029182626]
[Add reachable for data values
Neil Mitchell**20061029182610]
[Move to using records for all the Core types
Neil Mitchell**20061029181727]
[Remove special treatment for Prelude.[] and Prelude.: only in the show interface, this behaviour either wants to be put in the Convertor or dropped entirely (as it is now)
Neil Mitchell**20061029180414]
[Actually generate a completely linked file
Neil Mitchell**20061029175124]
[Write more of the linker, now actually does basic linking
Neil Mitchell**20061029172603]
[Pass around the free variables in a let binding, otherwise they are translated as functions
Neil Mitchell**20061029172521]
[Add Yhc.Core.Reachable
Neil Mitchell**20061029172439]
[Add some type signatures, to fix up the warnings
Neil Mitchell**20061029164532]
[Move to using records for the main Core type
Neil Mitchell**20061029164111]
[Add coreFuncName method
Neil Mitchell**20061029164007]
[Add some type signatures for Yhc.Core.Show, fix some warnings
Neil Mitchell**20061029142827]
[Use the trim functions from Util.Text instead of writing custom ones
Neil Mitchell**20061029142752]
[Add some type signatures to remove warnings
Neil Mitchell**20061029142726]
[core_js_step_0
golubovsky at gmail.com**20061029035154
Fixed a minor omission in Core.Convert to convert STRING primitive
to CoreStr.
Got Core to Javascript converter aligned with the latest changes in the Core
types. Got partially working string unmarshaller (works for string literals,
but not for char consed together).
]
[ycr2js_init_checkin
golubovsky at gmail.com**20061028032454
Code for ycr2js has been roken down into modules and transferred
into the Yhc project. Everything compiles but would not work at this time
until changes in Core haven't been addressed.
]
[Use the -core flag, used to be called -corep
Neil Mitchell**20061027194301]
[Add a basic interface to the linker, the linker doesn't do anything yet
Neil Mitchell**20061027180724]
[Add getNodes method to graph, required for Core link mode
Neil Mitchell**20061027180557]
[Add a -linkcore flag, currently unimplemented
Neil Mitchell**20061027174530]
[Change the Core flags, so -core generates a file and -showcore shows the file instead of generating it - also fixed the flags reference
Neil Mitchell**20061027170810]
[Make a file dirty if the flag -corep is specified and the core file is dirty
Neil Mitchell**20061027170141]
[Add an entry for .ycr files into the file flags, to make .ycr output more standard and less hacky
Neil Mitchell**20061027170050]
[Add CorePrim to the Core library, for primitives (allows the linking to be complete in some sense)
Neil Mitchell**20061027162324]
[Add a failure variable to the Core conversion, and make sure its used - fixes a bug with nested fatbar's and multiple replacements
Neil Mitchell**20061027154627]
[Move the code around for the CoreExpr conversion stuff, to make it a bit clearer
Neil Mitchell**20061027152357]
[Rename Core.Core to Core.Convert, since its the convertor, and no longer does all the Core stuff
Neil Mitchell**20061027145248]
[fix_version_py
golubovsky at gmail.com**20061027032708
version.py behaved incorrectly when _darcs/inventory's last line did not contain
two asterisks. Added a filter that keeps only lines with two asterisks.
]
[fix_core_show_and_init_js_checkin
golubovsky at gmail.com**20061027025815
Fixed Yhc.Core.Show to output CoreFloat and CoreDouble.
Initial directory structure and basic files check-in for
the Core -> Javascript converter.
]
[non_crashing_corefunc
golubovsky at gmail.com**20061026031128
Added a non-crashing version of coreFunc: returns Nothing when a function
is not found in the core.
]
[Remove a backup of the old wiki, out of date information (the new wiki is backed up automatically)
Neil Mitchell**20061026133230]
[Add a Fun type in addition to Var, to indicate a top level binding - useful for reachability analysis
Neil Mitchell**20061025175845]
[Remove the unused function argument from exp2core
Neil Mitchell**20061025173329]
[Give concatMapM a type signature
Neil Mitchell**20061025152529]
[Redefine concatMapM as a liftM
Neil Mitchell**20061025152434]
[Use a proper pretty printer for Yhc.Core
Neil Mitchell**20061025152043]
[Generalise the getHaskellFile, to allow yhc core files to be given without arguments
Neil Mitchell**20061025145735]
[Add core viewer, -viewcore, to the Yhc
Neil Mitchell**20061025144731]
[Fix isTidOp, so that an identifier only of ___ works
Neil Mitchell**20061025142641]
[Build Yhc.Core library
Andrew Wilkinson <aw at cs.york.ac.uk>**20061025141839]
[Move the Yhc.Core into a separate libraries directory
Neil Mitchell**20061025135804]
[Fix so that building without cpp gives you (unversioned) and still works
Neil Mitchell**20061025135031]
[Rename the version variable.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061025132216]
[Make the new Yhc Libraries build
Andrew Wilkinson <aw at cs.york.ac.uk>**20061025131852]
[Create a libraries directory, and start on a Yhc.General library - will not compile!
Neil Mitchell**20061025125517]
[Switch to using date/time of last patch as version number to make it consistant across all computers.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061025111200]
[Build and install .ycr files for Haskell libraries.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061025110529]
[Don't try and build files unless we really need them.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061022173843]
[Fix yhc build error on Windows buildbot.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061021175249]
[Python people have moved the directory for libffi.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061021172116]
[Fix building the docs now yhe can be built.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061021170442]
[Add support for building Yhe.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061021143044]
[Fix building the docs.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061020135037]
[Add file with version number in it.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061020133953]
[Add support for proper version numbers.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061020133417]
[Fix Samuel Bronson's broken patch. All tests pass with this - not sure if it's the correct fix though.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061020110049]
[Keep generated names from conflicting with real names
Samuel Bronson <naesten at gmail.com>**20061019200517]
[Fix more Id output stuff
Samuel Bronson <naesten at gmail.com>**20061019194125]
[Allow user to specify addition directories to search for libraries.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061016160752]
[Sigs and Haddock
Samuel Bronson <naesten at gmail.com>**20061016000836]
[Revert Show instance for Id
Samuel Bronson <naesten at gmail.com>**20061016000304]
[Fix a bug, Id's were being shown differently since being newtype wrapped
Neil Mitchell**20061014201710]
[Add version check for Darcs.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061013160337]
[Remove old ghc file.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061013104525]
[Check for a recent version of Python.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061013104346]
[Fix scons depends on Windows. (Bug #66)
Andrew Wilkinson <aw at cs.york.ac.uk>**20061013081625]
[Fix building yhi on Windows.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061012160123]
[Hack around with the linker flags to allow GMP to link.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061012154228]
[Don't use a blank environment to the always_configure.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061012152614]
[Allow libgmp to be called GMP.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061012152225]
[Rewrite the readme, old version was just plain wrong
Neil Mitchell**20061012151059]
[Look for libgmp where GHC installs it on Mac OS X.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061012150636]
[Make basepath_guess try adding a .exe on the end if a filename cannot be located, fixes bug on Windows
Neil Mitchell**20061008174601]
[Free the path on basepath_guess
Neil Mitchell**20061008172222]
[Make sure that the last entry on the path is searched, by appending a SEPARATOR character on to the end
Neil Mitchell**20061008171825]
[Add lexing change for 1e12 as a rational, from Malcolm, extracted from nhc
Neil Mitchell**20061008160957]
[Add tests for parsing 1e12, common bug for nhc
Neil Mitchell**20061008160744]
[Add type signatures to the test program
Neil Mitchell**20061008155549]
[Add scons push.
Andrew Wilkinson <aw at cs.york.ac.uk>**20061008152745]
[Types in Type.Type
Samuel Bronson <naesten at gmail.com>**20061007170726]
[TAG 07_oct_2006
Tom Shackell <shackell at cs.york.ac.uk>**20061007165033]
Patch bundle hash:
3787437454e64476f139adaa1a41e2e185b0be63
More information about the Yhc
mailing list