August 2007 - Yhc - Haskell.org

buildbot failure in macosx_ppc_3
by andrewjwilkinsonw＠gmail.com 29 Aug '07

29 Aug '07

The Buildbot has detected a new failure of macosx_ppc_3. Full details are available at: http://www.indiegigs.co.uk:8010/macosx_ppc_3/builds/139 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: macosxppc3 Build Reason: Build Source Stamp: HEAD Blamelist: golubovsky(a)gmail.com BUILD FAILED: failed failed slave lost sincerely, -The Buildbot

1 0

Config error..PreludeBuiltin...is this still in the build?
by Alexis Morris 29 Aug '07

29 Aug '07

Hi all, I just built a fresh yhc/yhi (after only adding one line to Primitive.hs) and everything seemed fine. I then copied my inst folder to c:\yhcinstall\. When I run "yhc hello.hs" however I get the following: $ yhc hello.hs Error: File not found, PreludeBuiltin Reason: imported from Prelude Looked in: c:\Documents and Settings\Alexis\Desktop\2007\NXHaskell\current\BCKernel\ c:\yhcinstall\lib\yhc\packages\haskell98\1.0 c:\yhcinstall\lib\yhc\packages\yhc-base\1.0 I'm not sure if this is a bug in the config or is caused by copying. I do have a PreludeBuiltin in my C:/YHC_BASE/version220807/yhc/src/packages/yhc-base-1.0/PreludeBuiltin.hi <file:///C:/YHC_BASE/version220807/yhc/src/packages/yhc-base-1.0/PreludeBuiltin.hi> I'll see if copying it over helps....nope...copying it seems to cause yhc loop until stack overflow. $ yhc hello.hs Stack space overflow: current size 8388608 bytes. Use `+RTS -Ksize' to increase it. Cheers, -- </Alexis>

2 3

Adding Custom Primitive functions
by Alexis Morris 29 Aug '07

29 Aug '07

Hi all, I have a number of driver related function calls that I would like to add as Primitive functions to yhc and yhi. For yhi (yhc\src\runtime\BCKernel) I have added definitions to Runtime.c and WrapPrimitive.c, in order to define and wrap stuff like "void display_string(void)" where the function is declared in a display.h/display.c file (this code controls the LCD display of the NXT). So far I think I have this yhi part covered. My question is about where these primitives are made known to the compiler so I can make display_string visible to yhc programs. After adding this line to Primitive.hs yhc hangs... ---------------------------------------------------------------------------------------------------- -- NXT Display ---------------------------------------------------------------------------------------------------- foreign import fastccall display_string :: CString -> IO () I just noticed the following files seem to have something todo with things: C:\YHC_OLD_SOURCES\yhc_again\yhc\src\packages\yhc-base-1.0\YHCPrimitive.hs C:\YHC_OLD_SOURCES\yhc_again\yhc\src\compiler\Prim.hs C:\YHC_OLD_SOURCES\yhc_again\yhc\src\compiler\PrimCode.hs C:\YHC_OLD_SOURCES\yhc_again\yhc\src\compiler98\Prim.hs C:\YHC_OLD_SOURCES\yhc_again\yhc\src\compiler98\PrimCode.hs Which changes I need to make before recompiling YHC? Thanks in advance. :) Cheers, -- </Alexis>

2 1

YHI Loader ...generated bytecode as C but some errors.
by Alexis Morris 29 Aug '07

29 Aug '07

Hi Tom, Thanks for your detailed instructions about splitting YHI into two parts and then loading bytecode from the heap. I followed them, and was able to generate a file to be loaded into yhi. However, upon compiling I get some nasty errors that I'm unable to figure out. Below is a snippet from a factorial function: fac 0 = 1 fac n = n * fac (n-1) main = print (fac 42) -------------------------------------- //this generates....among other things -------------------------------------- . . .snip... UInt8 bytecode_Main_2Emain[] = { 0X1, 0X25, 0X26, 0X8, 0X31, 0X42, 0X1, 0X8, 0X32, 0XC, 0X33, 0X34, 0X5B, 0 }; UByte ctypes_Main_2Emain[] = { C_NODE, C_NODE, C_INFO, C_INFO, C_INFO, C_INFO }; PInfo pinfo_Main_2Emain[] ={ }; FInfo finfo_Main_2Emain = (FInfo){ {I_FINFO}, pinfo_Main_2Emain, NULL, 0, 3, 0x00, NULL, NULL, 14, &bytecode_Main_2Emain, 6, &ctypes_Main_2Emain, &ctable_Main_2Emain }; ConstItem ctable_Main_2Emain[] = { (ConstItem) &node_0x18720028, (ConstItem) &node_0x18720480, (ConstItem) &finfo_Main_2Emain, (ConstItem) &finfo_Main_2Emain, (ConstItem) &finfo_Main_2Emain, (ConstItem) &finfo_Main_2Emain }; Word node_0x18720024[] = { (Word) &pinfo_Main_2Emain[0] }; UInt8 bytecode_Main_2Efac[] = { 0X1, 0X25, 0X12, 0X30, 0X42, 0X1, 0XB, 0X13, 0X12, 0X31, 0X32, 0X5A, 0X1, 0X41, 0X2, 0X1, 0X5A, 0X1, 0X4D, 0X5F, 0XB, 0, 0X1, 0X29, 0X17, 0X30, 0X5A, 0X1, 0X42, 0X1, 0X5B, 0X1, 0X29, 0X12, 0X30, 0X42, 0X1, 0X13, 0X12, 0X34, 0X41, 0X2, 0X1, 0X12, 0X35, 0XB, 0X18, 0X17, 0X36, 0X5A, 0X1, 0X41, 0X2, 0X1, 0X5B, 0 }; UByte ctypes_Main_2Efac[] = { C_NODE, C_INFO, C_INFO, C_INFO, C_NODE, C_INFO, C_INFO, C_INFO }; PInfo pinfo_Main_2Efac[] ={ (PInfo){ { I_PINFO }, 0, 2 }, (PInfo){ { I_PINFO }, 1, 1 } }; FInfo finfo_Main_2Efac = (FInfo){ {I_FINFO}, pinfo_Main_2Efac, NULL, 2, 4, 0x00, NULL, NULL, 56, &bytecode_Main_2Efac, 8, &ctypes_Main_2Efac, &ctable_Main_2Efac }; ConstItem ctable_Main_2Efac[] = { (ConstItem) &node_0x18720494, (ConstItem) &finfo_Main_2Efac, (ConstItem) &finfo_Main_2Efac, (ConstItem) &finfo_Main_2Efac, (ConstItem) &node_0x187204a4, (ConstItem) &finfo_Main_2Efac, (ConstItem) &finfo_Main_2Efac, (ConstItem) &finfo_Main_2Efac }; Word node_0x18720490[] = { (Word) &pinfo_Main_2Efac[0] }; ----------------------------------------------------------------------- When I compile I get these errors for this section of code ../yhi/loaderCode.c:15520: error: 'ctable_Main_2Emain' undeclared here (not in a function) ../yhi/loaderCode.c:15520: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15520: error: initializer element is not constant ../yhi/loaderCode.c:15520: error: (near initialization for 'finfo_Main_2Emain') ../yhi/loaderCode.c:15520: error: initializer element is not constant ../yhi/loaderCode.c:15524: error: 'node_0x18720480' undeclared here (not in a function) ../yhi/loaderCode.c:15524: error: initializer element is not constant ../yhi/loaderCode.c:15524: error: (near initialization for 'ctable_Main_2Emain[1]') ../yhi/loaderCode.c:15556: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15558: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15560: error: 'ctable_Main_2Efac' undeclared here (not in a function) ../yhi/loaderCode.c:15560: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15560: error: initializer element is not constant ../yhi/loaderCode.c:15560: error: (near initialization for 'finfo_Main_2Efac') ../yhi/loaderCode.c:15560: error: initializer element is not constant ../yhi/loaderCode.c:15563: error: 'node_0x18720494' undeclared here (not in a function) ../yhi/loaderCode.c:15563: error: initializer element is not constant ../yhi/loaderCode.c:15563: error: (near initialization for 'ctable_Main_2Efac[0]') ../yhi/loaderCode.c:15567: error: 'node_0x187204a4' undeclared here (not in a function) ../yhi/loaderCode.c:15567: error: initializer element is not constant ../yhi/loaderCode.c:15567: error: (near initialization for 'ctable_Main_2Efac[4]') ../yhi/loaderCode.c:15597: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15599: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15601: error: 'ctable_Data_2FChar_2ELAMBDA1036' undeclared here (not in a function) ../yhi/loaderCode.c:15601: warning: initialization from incompatible pointer type ../yhi/loaderCode.c:15601: error: initializer element is not constant ../yhi/loaderCode.c:15601: error: (near initialization for 'finfo_Data_2FChar_2ELAMBDA1036') ../yhi/loaderCode.c:15601: error: initializer element is not constant ../yhi/loaderCode.c:15605: error: 'node_0x1872038c' undeclared here (not in a function) ------------------------------------------------------------ Any ideas about why that blows up? In particular I'm curious about this error, "error: initializer element is not constant" not sure what it means. Thanks for the help, -- </Alexis>

2 1

Crashing Queens
by Neil Mitchell 29 Aug '07

29 Aug '07

Hi, Using the attached program: -------------------------------------------------------------------------------- D:\Temp>yhc Queens.hs Compiling Queens ( Queens.hs ) D:\Temp>yhi Queens.hbc Assertion failed: pinfo->info.tag == I_PINFO, file d:\sources\yhc\current\src\ru ntime\bckernel\mutins.h, line 295 This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information. -------------------------------------------------------------------------------- This also shows we have assertions on in the release "scons yhi" code, I guess they can stay in for now, but we really do need an optimisation mode. Program from Matt :-) Thanks Neil ----------- Queens.hs -------------- module Queens where main = nsoln 9 len :: [a] -> Int len [] = 0 len (x:xs) = 1 + len xs nsoln :: Int -> Int nsoln nq = len (gen nq) where gen :: Int -> [[Int]] gen 0 = [[]] gen n = [ (q:b) | b <- gen (n-1), q <- [1,2,3,4,5,6,7,8,9] {-toOne nq-}, safe q 1 b] safe :: Int -> Int -> [Int] -> Bool safe x d [] = True safe x d (q:l) = x /= q && x /= q+d && x /= q-d && safe x (d+1) l toOne :: Int -> [Int] toOne n = if n == 1 then [1] else n : toOne (n-1)

2 2

Adding new calling conventions for use in YHC.Core
by Dimitry Golubovsky 22 Aug '07

22 Aug '07

Jeremy, You wrote: ======================== Perhaps when the --core flag is present, the frontend should just treat the calling convention as String? Basically, I would extend the Style from: data Style = Ordinary | CCast | Address | FunAddress | Dynamic | Wrapper deriving Eq to: data Style = Ordinary | CCast | Address | FunAddress | Dynamic | Wrapper | Other String deriving Eq ======================== This message: http://www.haskell.org/pipermail/yhc/2007-August/001083.html mentions this (in the last paragraph). So, was this actually implemented, or still remains a proposal? -- Dimitry Golubovsky Anywhere on the Web

2 1

Adding new calling conventions for use in YHC.Core
by Jeremy Shaw 22 Aug '07

22 Aug '07

Hello, My ycr->swf backend is far enough along that I am ready to figure out how to properly import foreign functions. There are three types of ActionScript 'functions' that I want to import: 1. actionscript primitives, such as ActionSubtract, ActionStringAdd, etc 2. actionscript functions (mostly old functions that predate flash 5.0) 3. actionscript method calls (for flash 5.0 and higher, which has an object system) These each require a slightly different calling convention. So, I think I want to be able to import them by adding lines like this: foreign import primitive "ActionSubtract" actionSubtractInt :: Int -> Int -> Int foreign import ascall "someFunction" someFunction :: Int -> IO Int foreign import asmethod "someMethod" someMethod :: SomeObject -> Int -> IO Int However, to do this requires modifying the yhc frontend to accept these additional calling conventions. In YHC.Core, the calling convention is just a String, so there is no problem representing my new calling conventions in Core. But, the YHC frontend only wants to accept a limited set of predefined calling conventions. Perhaps when the --core flag is present, the frontend should just treat the calling convention as String? Basically, I would extend the Style from: data Style = Ordinary | CCast | Address | FunAddress | Dynamic | Wrapper deriving Eq to: data Style = Ordinary | CCast | Address | FunAddress | Dynamic | Wrapper | Other String deriving Eq And then modify the rest of the code so that when yhc is compiled with the --core flag, the calling convention is not parsed, but simply stuck inside the Other constructor? (Or perhaps, if it parses, the normal constructor is used, but if it fails to parse, then it is just dumped in Other ?) Or, perhaps I am taking completely the wrong approach ? j.

1 0

Possible bug in Yhc bytecode generator
by Dimitry Golubovsky 22 Aug '07

22 Aug '07

Hi, This file: http://www.golubovsky.org/repos/FudgetsOnWeb/Graphics/UI/Fudgets/SP/Spops.hs caused the following error message during compilation: Compiling Graphics.UI.Fudgets.SP.Spops ( ../FudgetsOnWeb/Graphics/UI/Fudgets/SP/Spops.hs ) yhc: src/compiler/ByteCode/CompileYhcCore.hs:(228,4)-(249,20): Non-exhaustive patterns in case This error did not occur when compiling with the --no-bytecode option Thanks. -- Dimitry Golubovsky Anywhere on the Web

1 0

Update on Javascript backend
by Dimitry Golubovsky 19 Aug '07

19 Aug '07

Hi, I have committed patches to fix incompatibilities in the Javascript backend that resulted from recent canges in Yhc Core. The EchoCPS demo now compiles and works. Anybody interested in experimenting with the Javascript backend are welcome to test the new code. Thanks. -- Dimitry Golubovsky Anywhere on the Web

1 0

buildbot failure in windows
by andrewjwilkinsonw＠gmail.com 15 Aug '07

15 Aug '07

The Buildbot has detected a new failure of windows. Full details are available at: http://www.indiegigs.co.uk:8010/windows/builds/182 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: neilwindows Build Reason: Build Source Stamp: HEAD Blamelist: Neil Mitchell BUILD FAILED: failed failed slave lost sincerely, -The Buildbot

1 0