From 3fda793f7428ebfdd595c344ee7f3d60fca1f996 Mon Sep 17 00:00:00 2001 From: Eugen Wissner Date: Thu, 15 Feb 2024 15:13:47 +0100 Subject: [PATCH] Start a Modula-2 experiment --- .gitignore | 2 +- README | 37 - README.md | 63 ++ Rakefile | 9 +- compiler.mod | 522 ++++++++++++ dub.json | 9 - source/elna/arguments.d | 154 ---- source/elna/backend.d | 100 --- source/elna/elf.d | 1060 ------------------------ source/elna/extended.d | 335 -------- source/elna/ir.d | 272 ------ source/elna/lexer.d | 254 ------ source/elna/parser.d | 372 --------- source/elna/result.d | 107 --- source/elna/riscv.d | 352 -------- source/main.d | 33 - tests/const_list.eln | 3 - tests/expectations/const_list.txt | 1 - tests/expectations/left_nested_sum.txt | 1 - tests/expectations/subtraction.txt | 1 - tests/expectations/sum.txt | 1 - tests/expectations/sums.txt | 1 - tests/left_nested_sum.eln | 2 - tests/subtraction.eln | 2 - tests/sum.eln | 2 - tests/sums.eln | 2 - 26 files changed, 591 insertions(+), 3106 deletions(-) delete mode 100644 README create mode 100644 README.md create mode 100644 compiler.mod delete mode 100644 dub.json delete mode 100644 source/elna/arguments.d delete mode 100644 source/elna/backend.d delete mode 100644 source/elna/elf.d delete mode 100644 source/elna/extended.d delete mode 100644 source/elna/ir.d delete mode 100644 source/elna/lexer.d delete mode 100644 source/elna/parser.d delete mode 100644 source/elna/result.d delete mode 100644 source/elna/riscv.d delete mode 100644 source/main.d delete mode 100644 tests/const_list.eln delete mode 100644 tests/expectations/const_list.txt delete mode 100644 tests/expectations/left_nested_sum.txt delete mode 100644 tests/expectations/subtraction.txt delete mode 100644 tests/expectations/sum.txt delete mode 100644 tests/expectations/sums.txt delete mode 100644 tests/left_nested_sum.eln delete mode 100644 tests/subtraction.eln delete mode 100644 tests/sum.eln delete mode 100644 tests/sums.eln diff --git a/.gitignore b/.gitignore index d0d201a..cde7e11 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,3 @@ -/.dub/ +a.out /dub.selections.json /build/ diff --git a/README b/README deleted file mode 100644 index b6e03eb..0000000 --- a/README +++ /dev/null @@ -1,37 +0,0 @@ -# Elna programming language - -Elna compiles simple mathematical operations to machine code. -The compiled program returns the result of the operation. - -## File extension - -.elna - -## Grammar PL/0 - -program = block "." ; - -block = [ "const" ident "=" number {"," ident "=" number} ";"] - [ "var" ident {"," ident} ";"] - { "procedure" ident ";" block ";" } statement ; - -statement = [ ident ":=" expression | "call" ident - | "?" ident | "!" expression - | "begin" statement {";" statement } "end" - | "if" condition "then" statement - | "while" condition "do" statement ]; - -condition = "odd" expression | - expression ("="|"#"|"<"|"<="|">"|">=") expression ; - -expression = [ "+"|"-"] term { ("+"|"-") term}; - -term = factor {("*"|"/") factor}; - -factor = ident | number | "(" expression ")"; - -## Operations - -"!" - Write a line. -"?" - Read user input. -"odd" - The only function, returns whether a number is odd. diff --git a/README.md b/README.md new file mode 100644 index 0000000..211ca55 --- /dev/null +++ b/README.md @@ -0,0 +1,63 @@ +# Elna programming language + +Elna is a simple, imperative, low-level programming language. + +It is intendet to accompany other languages in the areas, where a high-level +language doesn't fit well. It is also supposed to be an intermediate +representation for a such high-level hypothetical programming language. + +## File extension + +.elna + +## Current implementation + +This repository contains a GCC frontend for Elna. After finishing the frontend +I'm planning to rewrite the compiler in Elna itself with its own backend and +a hand-written parser. So GCC gives a way to have a simple bootstrap compiler +and a possbility to compile Elna programs for different platforms. + +## Grammar + +program = block "." ; + +block = [ "const" ident "=" number {"," ident "=" number} ";"] + [ "var" ident {"," ident} ";"] + { "procedure" ident ";" block ";" } statement ; + +statement = [ ident ":=" expression | "call" ident + | "?" ident | "!" expression + | "begin" statement {";" statement } "end" + | "if" condition "then" statement + | "while" condition "do" statement ]; + +condition = "odd" expression | + expression ("="|"#"|"<"|"<="|">"|">=") expression ; + +expression = [ "+"|"-"] term { ("+"|"-") term}; + +term = factor {("*"|"/") factor}; + +factor = ident | number | "(" expression ")"; + +## Build + +The frontend requires GCC 14.2.0 (not tested with other versions). + +Download the GCC source. Copy the contents of this repository into `gcc/elna` +inside GCC. Finally build GCC enabling the frontend with +`--enable-languages=c,c++,elna`. After the installation the compiler can be +invoked with `$prefix/bin/gelna`. + +There is also a `Rakefile` that downloads, builds and installs GCC into the +`./build/` subdirectory. The `Rakefile` assumes that ruby and rake, as well as +all GCC dependencies are already available in the system. It works under Linux +and Mac OS. In the latter case GCC is patched with the patches used by Homebrew +(official GCC doesn't support Apple silicon targets). Invoke with + +```sh +rake boot +``` + +See `rake -T` for more tasks. The GCC source is under `build/tools`. The +installation path is `build/host/install`. diff --git a/Rakefile b/Rakefile index 908dfdf..025ed5e 100644 --- a/Rakefile +++ b/Rakefile @@ -21,13 +21,13 @@ CLEAN.include '.dub' rule(/build\/riscv\/[^\/\.]+$/ => ->(file) { test_for_out(file, '.o') }) do |t| sh '/opt/riscv/bin/riscv32-unknown-elf-ld', '-o', t.name, - '-L/opt/riscv/lib/gcc/riscv32-unknown-elf/11.1.0', + '-L/opt/riscv/lib/gcc/riscv32-unknown-elf/13.2.0/', '-L/opt/riscv/riscv32-unknown-elf/lib', '/opt/riscv/riscv32-unknown-elf/lib/crt0.o', - '/opt/riscv/lib/gcc/riscv32-unknown-elf/11.1.0/crtbegin.o', + '/opt/riscv/lib/gcc/riscv32-unknown-elf/13.2.0/crtbegin.o', t.source, '--start-group', '-lgcc', '-lc', '-lgloss', '--end-group', - '/opt/riscv/lib/gcc/riscv32-unknown-elf/11.1.0/crtend.o' + '/opt/riscv/lib/gcc/riscv32-unknown-elf/13.2.0/crtend.o' end rule(/build\/riscv\/.+\.o$/ => ->(file) { test_for_object(file, '.eln') }) do |t| @@ -36,7 +36,7 @@ rule(/build\/riscv\/.+\.o$/ => ->(file) { test_for_object(file, '.eln') }) do |t end file BINARY => SOURCES do |t| - sh({ 'DFLAGS' => (DFLAGS * ' ') }, 'dub', 'build', '--compiler=gdc-12') + sh({ 'DFLAGS' => (DFLAGS * ' ') }, 'dub', 'build', '--compiler=gdc') end task default: BINARY @@ -55,6 +55,7 @@ task test: BINARY do if test.include? '/riscv/' spike = [ '/opt/riscv/bin/spike', + '--isa=RV32IMAC', '/opt/riscv/riscv32-unknown-elf/bin/pk', test ] diff --git a/compiler.mod b/compiler.mod new file mode 100644 index 0000000..4df7d4f --- /dev/null +++ b/compiler.mod @@ -0,0 +1,522 @@ +MODULE Compiler; + +FROM Terminal IMPORT WriteString, WriteLn; + +TYPE + (* + Classification table assigns each possible character to a group (class). All + characters of the same group a handled equivalently. + + Classification: + *) + TransitionClass = ( + transitionClassInvalid, + transitionClassDigit, + transitionClassCharacter, + transitionClassSpace, + transitionClassColon, + transitionClassEquals, + transitionClassLeftParen, + transitionClassRightParen, + transitionClassAsterisk, + transitionClassUnderscore, + transitionClassSingle, + transitionClassHex, + transitionClassZero, + transitionClassX, + transitionClassEof, + transitionClassDot, + transitionClassMinus, + transitionClassQuote, + transitionClassGreater, + transitionClassLess, + transitionClassOther + ); + TransitionState = ( + transitionStateStart, + transitionStateColon, + transitionStateIdentifier, + transitionStateDecimal, + transitionStateGreater, + transitionStateMinus, + transitionStateLeftParen, + transitionStateLess, + transitionStateHexadecimal, + transitionStateComment, + transitionStateClosingComment, + transitionStateString, + transitionStateLeadingZero, + transitionStateHexadecimalPrefix, + transitionStateEnd + ); + TransitionAction = PROCEDURE(); + Transition = RECORD + Action: TransitionAction; + NextState: TransitionState + END; + +VAR + Classification: ARRAY[1..128] OF TransitionClass; + Transitions: ARRAY[0..MAX(TransitionState)] OF ARRAY[0..MAX(TransitionClass)] OF Transition; + +PROCEDURE InitializeClassification(); +BEGIN + Classification[1] := transitionClassEof; (* NUL *) + Classification[2] := transitionClassInvalid; (* SOH *) + Classification[3] := transitionClassInvalid; (* STX *) + Classification[4] := transitionClassInvalid; (* ETX *) + Classification[5] := transitionClassInvalid; (* EOT *) + Classification[6] := transitionClassInvalid; (* EMQ *) + Classification[7] := transitionClassInvalid; (* ACK *) + Classification[8] := transitionClassInvalid; (* BEL *) + Classification[9] := transitionClassInvalid; (* BS *) + Classification[10] := transitionClassSpace; (* HT *) + Classification[11] := transitionClassSpace; (* LF *) + Classification[12] := transitionClassInvalid; (* VT *) + Classification[13] := transitionClassInvalid; (* FF *) + Classification[14] := transitionClassSpace; (* CR *) + Classification[15] := transitionClassInvalid; (* SO *) + Classification[16] := transitionClassInvalid; (* SI *) + Classification[17] := transitionClassInvalid; (* DLE *) + Classification[18] := transitionClassInvalid; (* DC1 *) + Classification[19] := transitionClassInvalid; (* DC2 *) + Classification[20] := transitionClassInvalid; (* DC3 *) + Classification[21] := transitionClassInvalid; (* DC4 *) + Classification[22] := transitionClassInvalid; (* NAK *) + Classification[23] := transitionClassInvalid; (* SYN *) + Classification[24] := transitionClassInvalid; (* ETB *) + Classification[25] := transitionClassInvalid; (* CAN *) + Classification[26] := transitionClassInvalid; (* EM *) + Classification[27] := transitionClassInvalid; (* SUB *) + Classification[28] := transitionClassInvalid; (* ESC *) + Classification[29] := transitionClassInvalid; (* FS *) + Classification[30] := transitionClassInvalid; (* GS *) + Classification[31] := transitionClassInvalid; (* RS *) + Classification[32] := transitionClassInvalid; (* US *) + Classification[33] := transitionClassSpace; (* Space *) + Classification[34] := transitionClassSingle; (* ! *) + Classification[35] := transitionClassQuote; (* " *) + Classification[36] := transitionClassOther; (* # *) + Classification[37] := transitionClassOther; (* $ *) + Classification[38] := transitionClassSingle; (* % *) + Classification[39] := transitionClassSingle; (* & *) + Classification[40] := transitionClassQuote; (* ' *) + Classification[41] := transitionClassLeftParen; (* ( *) + Classification[42] := transitionClassRightParen; (* ) *) + Classification[43] := transitionClassAsterisk; (* * *) + Classification[44] := transitionClassSingle; (* + *) + Classification[45] := transitionClassSingle; (* , *) + Classification[46] := transitionClassMinus; (* - *) + Classification[47] := transitionClassDot; (* . *) + Classification[48] := transitionClassSingle; (* / *) + Classification[49] := transitionClassZero; (* 0 *) + Classification[50] := transitionClassDigit; (* 1 *) + Classification[51] := transitionClassDigit; (* 2 *) + Classification[52] := transitionClassDigit; (* 3 *) + Classification[53] := transitionClassDigit; (* 4 *) + Classification[54] := transitionClassDigit; (* 5 *) + Classification[55] := transitionClassDigit; (* 6 *) + Classification[56] := transitionClassDigit; (* 7 *) + Classification[57] := transitionClassDigit; (* 8 *) + Classification[58] := transitionClassDigit; (* 9 *) + Classification[59] := transitionClassColon; (* : *) + Classification[60] := transitionClassSingle; (* ; *) + Classification[61] := transitionClassLess; (* < *) + Classification[62] := transitionClassEquals; (* = *) + Classification[63] := transitionClassGreater; (* > *) + Classification[64] := transitionClassOther; (* ? *) + Classification[65] := transitionClassSingle; (* @ *) + Classification[66] := transitionClassCharacter; (* A *) + Classification[67] := transitionClassCharacter; (* B *) + Classification[68] := transitionClassCharacter; (* C *) + Classification[69] := transitionClassCharacter; (* D *) + Classification[70] := transitionClassCharacter; (* E *) + Classification[71] := transitionClassCharacter; (* F *) + Classification[72] := transitionClassCharacter; (* G *) + Classification[73] := transitionClassCharacter; (* H *) + Classification[74] := transitionClassCharacter; (* I *) + Classification[75] := transitionClassCharacter; (* J *) + Classification[76] := transitionClassCharacter; (* K *) + Classification[77] := transitionClassCharacter; (* L *) + Classification[78] := transitionClassCharacter; (* M *) + Classification[79] := transitionClassCharacter; (* N *) + Classification[80] := transitionClassCharacter; (* O *) + Classification[81] := transitionClassCharacter; (* P *) + Classification[82] := transitionClassCharacter; (* Q *) + Classification[83] := transitionClassCharacter; (* R *) + Classification[84] := transitionClassCharacter; (* S *) + Classification[85] := transitionClassCharacter; (* T *) + Classification[86] := transitionClassCharacter; (* U *) + Classification[67] := transitionClassCharacter; (* V *) + Classification[88] := transitionClassCharacter; (* W *) + Classification[89] := transitionClassCharacter; (* X *) + Classification[90] := transitionClassCharacter; (* Y *) + Classification[91] := transitionClassCharacter; (* Z *) + Classification[92] := transitionClassSingle; (* [ *) + Classification[93] := transitionClassOther; (* \ *) + Classification[94] := transitionClassSingle; (* ] *) + Classification[95] := transitionClassSingle; (* ^ *) + Classification[96] := transitionClassUnderscore; (* _ *) + Classification[97] := transitionClassOther; (* ` *) + Classification[98] := transitionClassHex; (* a *) + Classification[99] := transitionClassHex; (* b *) + Classification[100] := transitionClassHex; (* c *) + Classification[101] := transitionClassHex; (* d *) + Classification[102] := transitionClassHex; (* e *) + Classification[103] := transitionClassHex; (* f *) + Classification[104] := transitionClassCharacter; (* g *) + Classification[105] := transitionClassCharacter; (* h *) + Classification[106] := transitionClassCharacter; (* i *) + Classification[107] := transitionClassCharacter; (* j *) + Classification[108] := transitionClassCharacter; (* k *) + Classification[109] := transitionClassCharacter; (* l *) + Classification[110] := transitionClassCharacter; (* m *) + Classification[111] := transitionClassCharacter; (* n *) + Classification[112] := transitionClassCharacter; (* o *) + Classification[113] := transitionClassCharacter; (* p *) + Classification[114] := transitionClassCharacter; (* q *) + Classification[115] := transitionClassCharacter; (* r *) + Classification[116] := transitionClassCharacter; (* s *) + Classification[117] := transitionClassCharacter; (* t *) + Classification[118] := transitionClassCharacter; (* u *) + Classification[119] := transitionClassCharacter; (* v *) + Classification[120] := transitionClassCharacter; (* w *) + Classification[121] := transitionClassX; (* x *) + Classification[122] := transitionClassCharacter; (* y *) + Classification[123] := transitionClassCharacter; (* z *) + Classification[124] := transitionClassOther; (* { *) + Classification[125] := transitionClassSingle; (* | *) + Classification[126] := transitionClassOther; (* } *) + Classification[127] := transitionClassSingle; (* ~ *) + Classification[128] := transitionClassInvalid (* DEL *) +END InitializeClassification; + +(* 0x00. No action. *) +PROCEDURE TransitionActionNo(); +BEGIN +END TransitionActionNo; + +(* 0x01. Accumulate action. *) +PROCEDURE TransitionActionAccumulate(); +BEGIN +END TransitionActionAccumulate; + +(* 0x02. Print action. *) +PROCEDURE TransitionActionPrint(); +BEGIN +END TransitionActionPrint; + +(* 0x03. Skip action. *) +PROCEDURE TransitionActionSkip(); +BEGIN +END TransitionActionSkip; + +(* 0x04. Delimited string action. *) +PROCEDURE TransitionActionDelimited(); +BEGIN +END TransitionActionDelimited; + +(* 0x05. Finalize identifier action. *) +PROCEDURE TransitionActionFinalize(); +BEGIN +END TransitionActionFinalize; + +(* 0x06. Single character symbol action. *) +PROCEDURE TransitionActionSingle(); +BEGIN +END TransitionActionSingle; + +(* 0x07. An action for symbols containing multiple characters. *) +PROCEDURE TransitionActionComposite(); +BEGIN +END TransitionActionComposite; + +(* 0x08. Integer action. *) +PROCEDURE TransitionActionInteger(); +BEGIN +END TransitionActionInteger; + +PROCEDURE SetDefaultTransition(currentState: TransitionState; DefaultAction: TransitionAction; NextState: TransitionState); +VAR DefaultTransition: Transition; +BEGIN + DefaultTransition.Action := DefaultAction; + DefaultTransition.NextState := NextState; + + Transitions[ORD(currentState)][ORD(transitionClassInvalid)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassDigit)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassCharacter)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassSpace)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassColon)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassEquals)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassLeftParen)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassRightParen)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassAsterisk)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassUnderscore)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassSingle)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassHex)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassZero)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassX)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassEof)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassDot)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassMinus)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassQuote)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassGreater)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassLess)] := DefaultTransition; + Transitions[ORD(currentState)][ORD(transitionClassOther)] := DefaultTransition; +END SetDefaultTransition; + +(* + * The transition table describes transitions from one state to another, given + * a symbol (character class). + * + * The table has m rows and n columns, where m is the amount of states and n is + * the amount of classes. So given the current state and a classified character + * the table can be used to look up the next state. + * + * Each cell is a word long. + * - The least significant byte of the word is a row number (beginning with 0). + * It specifies the target state. "ff" means that this is an end state and no + * transition is possible. + * - The next byte is the action that should be performed when transitioning. + * For the meaning of actions see labels in the lex_next function, which + * handles each action. + *) +PROCEDURE InitializeTransitions(); +BEGIN + (* Start state. *) + Transitions[ORD(transitionStateStart)][ORD(transitionClassInvalid)].Action := TransitionActionNo; + Transitions[ORD(transitionStateStart)][ORD(transitionClassInvalid)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassDigit)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassDigit)].NextState := transitionStateDecimal; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassCharacter)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassCharacter)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassSpace)].Action := TransitionActionSkip; + Transitions[ORD(transitionStateStart)][ORD(transitionClassSpace)].NextState := transitionStateStart; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassColon)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassColon)].NextState := transitionStateColon; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassEquals)].Action := TransitionActionSingle; + Transitions[ORD(transitionStateStart)][ORD(transitionClassEquals)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassLeftParen)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassLeftParen)].NextState := transitionStateLeftParen; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassRightParen)].Action := TransitionActionSingle; + Transitions[ORD(transitionStateStart)][ORD(transitionClassRightParen)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassAsterisk)].Action := TransitionActionSingle; + Transitions[ORD(transitionStateStart)][ORD(transitionClassAsterisk)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassUnderscore)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassUnderscore)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassSingle)].Action := TransitionActionSingle; + Transitions[ORD(transitionStateStart)][ORD(transitionClassSingle)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassHex)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassHex)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassZero)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassZero)].NextState := transitionStateLeadingZero; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassX)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassX)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassEof)].Action := TransitionActionNo; + Transitions[ORD(transitionStateStart)][ORD(transitionClassEof)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassDot)].Action := TransitionActionSingle; + Transitions[ORD(transitionStateStart)][ORD(transitionClassDot)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassMinus)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassMinus)].NextState := transitionStateMinus; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassQuote)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassQuote)].NextState := transitionStateString; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassGreater)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassGreater)].NextState := transitionStateGreater; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassLess)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateStart)][ORD(transitionClassLess)].NextState := transitionStateLess; + + Transitions[ORD(transitionStateStart)][ORD(transitionClassOther)].Action := TransitionActionNo; + Transitions[ORD(transitionStateStart)][ORD(transitionClassOther)].NextState := transitionStateEnd; + + (* Colon state. *) + SetDefaultTransition(transitionStateColon, TransitionActionPrint, transitionStateEnd); + + Transitions[ORD(transitionStateColon)][ORD(transitionClassEquals)].Action := TransitionActionComposite; + Transitions[ORD(transitionStateColon)][ORD(transitionClassEquals)].NextState := transitionStateEnd; + + (* Identifier state. *) + SetDefaultTransition(transitionStateIdentifier, TransitionActionFinalize, transitionStateEnd); + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassDigit)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassDigit)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassCharacter)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassCharacter)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassUnderscore)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassUnderscore)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassHex)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassHex)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassZero)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassZero)].NextState := transitionStateIdentifier; + + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassX)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateIdentifier)][ORD(transitionClassX)].NextState := transitionStateIdentifier; + + (* Decimal state. *) + SetDefaultTransition(transitionStateDecimal, TransitionActionInteger, transitionStateEnd); + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassDigit)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassDigit)].NextState := transitionStateDecimal; + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassCharacter)].Action := TransitionActionNo; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassCharacter)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassUnderscore)].Action := TransitionActionNo; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassUnderscore)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassHex)].Action := TransitionActionNo; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassHex)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassZero)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassZero)].NextState := transitionStateDecimal; + + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassX)].Action := TransitionActionNo; + Transitions[ORD(transitionStateDecimal)][ORD(transitionClassX)].NextState := transitionStateEnd; + + (* Greater state. *) + SetDefaultTransition(transitionStateGreater, TransitionActionPrint, transitionStateEnd); + + Transitions[ORD(transitionStateGreater)][ORD(transitionClassEquals)].Action := TransitionActionDelimited; + Transitions[ORD(transitionStateGreater)][ORD(transitionClassEquals)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateGreater)][ORD(transitionClassGreater)].Action := TransitionActionDelimited; + Transitions[ORD(transitionStateGreater)][ORD(transitionClassGreater)].NextState := transitionStateEnd; + + (* Minus state. *) + SetDefaultTransition(transitionStateGreater, TransitionActionSingle, transitionStateEnd); + + Transitions[ORD(transitionStateMinus)][ORD(transitionClassGreater)].Action := TransitionActionDelimited; + Transitions[ORD(transitionStateMinus)][ORD(transitionClassGreater)].NextState := transitionStateEnd; + + (* Left paren state. *) + SetDefaultTransition(transitionStateLeftParen, TransitionActionSingle, transitionStateEnd); + + Transitions[ORD(transitionStateLeftParen)][ORD(transitionClassAsterisk)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateLeftParen)][ORD(transitionClassAsterisk)].NextState := transitionStateComment; + + (* Less state. *) + SetDefaultTransition(transitionStateLess, TransitionActionSingle, transitionStateEnd); + + Transitions[ORD(transitionStateLess)][ORD(transitionClassLess)].Action := TransitionActionDelimited; + Transitions[ORD(transitionStateLess)][ORD(transitionClassLess)].NextState := transitionStateEnd; + + (* Hexadecimal after 0x. *) + SetDefaultTransition(transitionStateHexadecimal, TransitionActionInteger, transitionStateEnd); + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassDigit)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassDigit)].NextState := transitionStateHexadecimal; + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassCharacter)].Action := TransitionActionNo; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassCharacter)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassUnderscore)].Action := TransitionActionNo; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassUnderscore)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassHex)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassHex)].NextState := transitionStateHexadecimal; + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassZero)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassZero)].NextState := transitionStateHexadecimal; + + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassX)].Action := TransitionActionNo; + Transitions[ORD(transitionStateHexadecimal)][ORD(transitionClassX)].NextState := transitionStateEnd; + + (* Comment. *) + SetDefaultTransition(transitionStateComment, TransitionActionAccumulate, transitionStateComment); + + Transitions[ORD(transitionStateComment)][ORD(transitionClassAsterisk)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateComment)][ORD(transitionClassAsterisk)].NextState := transitionStateClosingComment; + + Transitions[ORD(transitionStateComment)][ORD(transitionClassEof)].Action := TransitionActionNo; + Transitions[ORD(transitionStateComment)][ORD(transitionClassEof)].NextState := transitionStateEnd; + + (* Closing comment. *) + SetDefaultTransition(transitionStateClosingComment, TransitionActionAccumulate, transitionStateComment); + + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassInvalid)].Action := TransitionActionNo; + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassInvalid)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassRightParen)].Action := TransitionActionDelimited; + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassRightParen)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassAsterisk)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassAsterisk)].NextState := transitionStateClosingComment; + + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassEof)].Action := TransitionActionNo; + Transitions[ORD(transitionStateClosingComment)][ORD(transitionClassEof)].NextState := transitionStateEnd; + + (* String. *) + SetDefaultTransition(transitionStateString, TransitionActionAccumulate, transitionStateString); + + Transitions[ORD(transitionStateString)][ORD(transitionClassInvalid)].Action := TransitionActionNo; + Transitions[ORD(transitionStateString)][ORD(transitionClassInvalid)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateString)][ORD(transitionClassEof)].Action := TransitionActionNo; + Transitions[ORD(transitionStateString)][ORD(transitionClassEof)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateString)][ORD(transitionClassQuote)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateString)][ORD(transitionClassQuote)].NextState := transitionStateEnd; + + (* Leading zero. *) + SetDefaultTransition(transitionStateLeadingZero, TransitionActionInteger, transitionStateEnd); + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassDigit)].Action := TransitionActionNo; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassDigit)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassCharacter)].Action := TransitionActionNo; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassCharacter)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassUnderscore)].Action := TransitionActionNo; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassUnderscore)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassHex)].Action := TransitionActionNo; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassHex)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassZero)].Action := TransitionActionNo; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassZero)].NextState := transitionStateEnd; + + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassX)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateLeadingZero)][ORD(transitionClassX)].NextState := transitionStateHexadecimalPrefix; + + (* Leading zero. *) + SetDefaultTransition(transitionStateHexadecimalPrefix, TransitionActionNo, transitionStateEnd); + + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassDigit)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassDigit)].NextState := transitionStateHexadecimal; + + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassHex)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassHex)].NextState := transitionStateHexadecimal; + + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassZero)].Action := TransitionActionAccumulate; + Transitions[ORD(transitionStateHexadecimalPrefix)][ORD(transitionClassZero)].NextState := transitionStateHexadecimal +END InitializeTransitions; + +BEGIN + InitializeClassification(); + InitializeTransitions(); + + WriteString('Elna'); + WriteLn() +END Compiler. diff --git a/dub.json b/dub.json deleted file mode 100644 index 8567d3a..0000000 --- a/dub.json +++ /dev/null @@ -1,9 +0,0 @@ -{ - "dependencies": { - "tanya": "~>0.18.0" - }, - "name": "elna", - "targetType": "executable", - "targetPath": "build/bin", - "mainSourceFile": "source/main.d" -} diff --git a/source/elna/arguments.d b/source/elna/arguments.d deleted file mode 100644 index 1ad928a..0000000 --- a/source/elna/arguments.d +++ /dev/null @@ -1,154 +0,0 @@ -/** - * Argument parsing. - */ -module elna.arguments; - -import std.algorithm; -import std.range; -import std.sumtype; - -struct ArgumentError -{ - enum Type - { - expectedOutputFile, - noInput, - superfluousArguments, - } - - private Type type_; - private string argument_; - - @property Type type() const @nogc nothrow pure @safe - { - return this.type_; - } - - @property string argument() const @nogc nothrow pure @safe - { - return this.argument_; - } - - void toString(OR)(OR range) - if (isOutputRage!OR) - { - final switch (Type) - { - case Type.expectedOutputFile: - put(range, "Expected an output filename after -o"); - break; - case Type.noInput: - put(range, "No input files specified"); - break; - } - } -} - -/** - * Supported compiler arguments. - */ -struct Arguments -{ - private bool assembler_; - private string output_; - private string inFile_; - - @property string inFile() @nogc nothrow pure @safe - { - return this.inFile_; - } - - /** - * Returns: Whether to generate assembly instead of an object file. - */ - @property bool assembler() const @nogc nothrow pure @safe - { - return this.assembler_; - } - - /** - * Returns: Output file. - */ - @property string output() const @nogc nothrow pure @safe - { - return this.output_; - } - - /** - * Parse command line arguments. - * - * The first argument is expected to be the program name (and it is - * ignored). - * - * Params: - * arguments = Command line arguments. - * - * Returns: Parsed arguments or an error. - */ - static SumType!(ArgumentError, Arguments) parse(string[] arguments) - @nogc nothrow pure @safe - { - if (!arguments.empty) - { - arguments.popFront; - } - alias ReturnType = typeof(return); - - return parseArguments(arguments).match!( - (Arguments parsed) { - if (parsed.inFile is null) - { - return ReturnType(ArgumentError(ArgumentError.Type.noInput)); - } - else if (!arguments.empty) - { - return ReturnType(ArgumentError( - ArgumentError.Type.superfluousArguments, - arguments.front - )); - } - return ReturnType(parsed); - }, - (ArgumentError argumentError) => ReturnType(argumentError) - ); - } - - private static SumType!(ArgumentError, Arguments) parseArguments(ref string[] arguments) - @nogc nothrow pure @safe - { - Arguments parsed; - - while (!arguments.empty) - { - if (arguments.front == "-s") - { - parsed.assembler_ = true; - } - else if (arguments.front == "-o") - { - if (arguments.empty) - { - return typeof(return)(ArgumentError( - ArgumentError.Type.expectedOutputFile, - arguments.front - )); - } - arguments.popFront; - parsed.output_ = arguments.front; - } - else if (arguments.front == "--") - { - arguments.popFront; - parsed.inFile_ = arguments.front; - arguments.popFront; - break; - } - else if (!arguments.front.startsWith("-")) - { - parsed.inFile_ = arguments.front; - } - arguments.popFront; - } - return typeof(return)(parsed); - } -} diff --git a/source/elna/backend.d b/source/elna/backend.d deleted file mode 100644 index 995efda..0000000 --- a/source/elna/backend.d +++ /dev/null @@ -1,100 +0,0 @@ -module elna.backend; - -import core.stdc.stdio; -import elna.elf; -import elna.ir; -import elna.extended; -import elna.riscv; -import elna.lexer; -import elna.parser; -import elna.result; -import std.algorithm; -import std.sumtype; -import std.typecons; -import tanya.os.error; -import tanya.container.array; -import tanya.container.string; -import tanya.memory.allocator; - -private Nullable!String readSource(string source) @nogc -{ - enum size_t bufferSize = 255; - auto sourceFilename = String(source); - - return readFile(sourceFilename).match!( - (ErrorCode errorCode) { - perror(sourceFilename.toStringz); - return Nullable!String(); - }, - (Array!ubyte contents) => nullable(String(cast(char[]) contents.get)) - ); -} - -int generate(string inFile, ref String outputFilename) @nogc -{ - auto sourceText = readSource(inFile); - if (sourceText.isNull) - { - return 3; - } - auto tokens = lex(sourceText.get.get); - if (!tokens.valid) - { - auto compileError = tokens.error.get; - printf("%lu:%lu: %s\n", compileError.line, compileError.column, compileError.message.ptr); - return 1; - } - auto ast = parse(tokens.result); - if (!ast.valid) - { - auto compileError = ast.error.get; - printf("%lu:%lu: %s\n", compileError.line, compileError.column, compileError.message.ptr); - return 2; - } - auto transformVisitor = defaultAllocator.make!TransformVisitor(); - auto ir = transformVisitor.visit(ast.result); - defaultAllocator.dispose(transformVisitor); - - auto handle = File.open(outputFilename.toStringz, BitFlags!(File.Mode)(File.Mode.truncate)); - if (!handle.valid) - { - return 1; - } - auto program = writeNext(ir); - auto elf = Elf!ELFCLASS32(move(handle)); - auto readOnlyData = Array!ubyte(cast(const(ubyte)[]) "%d\n".ptr[0 .. 4]); // With \0. - - elf.addReadOnlyData(String(".CL0"), readOnlyData); - elf.addCode(program.name, program.text); - - elf.addExternSymbol(String("printf")); - foreach (ref reference; program.symbols) - { - elf.Rela relocationEntry = { - r_offset: cast(elf.Addr) reference.offset - }; - elf.Rela relocationSub = { - r_offset: cast(elf.Addr) reference.offset, - r_info: R_RISCV_RELAX - }; - - final switch (reference.target) - { - case Reference.Target.text: - relocationEntry.r_info = R_RISCV_CALL; - break; - case Reference.Target.high20: - relocationEntry.r_info = R_RISCV_HI20; - break; - case Reference.Target.lower12i: - relocationEntry.r_info = R_RISCV_LO12_I; - break; - } - - elf.relocate(reference.name, relocationEntry, relocationSub); - } - - elf.finish(); - - return 0; -} diff --git a/source/elna/elf.d b/source/elna/elf.d deleted file mode 100644 index 81dd7da..0000000 --- a/source/elna/elf.d +++ /dev/null @@ -1,1060 +0,0 @@ -module elna.elf; - -import elna.extended; -import elna.result; -import std.algorithm; -import tanya.container.array; -import tanya.container.hashtable; -import tanya.container.string; - -/// Unsigned program address. -alias Elf64_Addr = ulong; -/// Unsigned file offset. -alias Elf64_Off = ulong; -/// Unsigned medium integer. -alias Elf64_Half = ushort; -/// Unsigned integer. -alias Elf64_Word = uint; -/// Signed integer. -alias Elf64_Sword = int; -/// Unsigned long integer. -alias Elf64_Xword = ulong; -/// Signed long integer. -alias Elf64_Sxword = long; - -/// Unsigned program address. -alias Elf32_Addr = uint; -/// Unsigned file offset. -alias Elf32_Off = uint; -/// Unsigned medium integer. -alias Elf32_Half = ushort; -/// Unsigned integer. -alias Elf32_Word = uint; -/// Signed integer. -alias Elf32_Sword = int; - -enum : size_t -{ - /// File identification. - EI_MAG0 = 0, - /// File identification. - EI_MAG1 = 1, - /// File identification. - EI_MAG2 = 2, - /// File identification. - EI_MAG3 = 3, - /// File class. - EI_CLASS = 4, - /// Data encoding. - EI_DATA = 5, - /// File version. - EI_VERSION = 6, - /// Start of padding bytes. - EI_PAD = 7, - /// Size of e_ident[] - EI_NIDENT = 16 -} - -enum : ubyte -{ - /// e_ident[EI_MAG0]. - ELFMAG0 = 0x7f, - /// e_ident[EI_MAG1]. - ELFMAG1 = 'E', - /// e_ident[EI_MAG2]. - ELFMAG2 = 'L', - /// e_ident[EI_MAG3]. - ELFMAG3 = 'F' -} - -/** - * File header. - */ -struct Elf64_Ehdr -{ - /// ELF identification. - ubyte[EI_NIDENT] e_ident; - /// Object file type. - Elf64_Half e_type; - /// Machine type. - Elf64_Half e_machine; - /// Object file version - Elf64_Word e_version; - /// Entry point address. - Elf64_Addr e_entry; - /// Program header offset. - Elf64_Off e_phoff; - /// Section header offset. - Elf64_Off e_shoff; - /// Processor-specific flags. - Elf64_Word e_flags; - /// ELF header size. - Elf64_Half e_ehsize; - /// Size of program header entry. - Elf64_Half e_phentsize; - /// Number of program header entries. - Elf64_Half e_phnum; - /// Size of section header entry. - Elf64_Half e_shentsize; - /// Number of section header entries. - Elf64_Half e_shnum; - /// Section name string table index. - Elf64_Half e_shstrndx; -} - -/** - * File header. - */ -struct Elf32_Ehdr { - /// ELF identification. - ubyte[EI_NIDENT] e_ident; - /// Object file type. - Elf32_Half e_type; - /// Machine type. - Elf32_Half e_machine; - /// Object file version - Elf32_Word e_version; - /// Entry point address. - Elf32_Addr e_entry; - /// Program header offset. - Elf32_Off e_phoff; - /// Section header offset. - Elf32_Off e_shoff; - /// Processor-specific flags. - Elf32_Word e_flags; - /// ELF header size. - Elf32_Half e_ehsize; - /// Size of program header entry. - Elf32_Half e_phentsize; - /// Number of program header entries. - Elf32_Half e_phnum; - /// Size of section header entry. - Elf32_Half e_shentsize; - /// Number of section header entries. - Elf32_Half e_shnum; - /// Section name string table index. - Elf32_Half e_shstrndx; -} - -/** - * Section header. - */ -struct Elf64_Shdr -{ - /// Section name. - Elf64_Word sh_name; - /// Section type. - Elf64_Word sh_type; - /// Section attributes. - Elf64_Xword sh_flags; - /// Virtual address in memory. - Elf64_Addr sh_addr; - /// Offset in file. - Elf64_Off sh_offset; - /// Size of section. - Elf64_Xword sh_size; - /// Link to other section. - Elf64_Word sh_link; - /// Miscellaneous information. - Elf64_Word sh_info; - /// Address alignment boundary. - Elf64_Xword sh_addralign; - /// Size of entries, if section has table. - Elf64_Xword sh_entsize; -} - -/** - * Section header. - */ -struct Elf32_Shdr -{ - /// Section name. - Elf32_Word sh_name; - /// Section type. - Elf32_Word sh_type; - /// Section attributes. - Elf32_Word sh_flags; - /// Virtual address in memory. - Elf32_Addr sh_addr; - /// Offset in file. - Elf32_Off sh_offset; - /// Size of section. - Elf32_Word sh_size; - /// Link to other section. - Elf32_Word sh_link; - /// Miscellaneous information. - Elf32_Word sh_info; - /// Address alignment boundary. - Elf32_Word sh_addralign; - /// Size of entries, if section has table. - Elf32_Word sh_entsize; -} - -/** - * Symbol table entry. - */ -struct Elf64_Sym -{ - /// Symbol name. - Elf64_Word st_name; - /// Type and Binding attributes. - ubyte st_info; - /// Reserved. - ubyte st_other; - /// Section table index. - Elf64_Half st_shndx; - /// Symbol value. - Elf64_Addr st_value; - /// Size of object (e.g., common). - Elf64_Xword st_size; -} - -/** - * Relocation entry. - */ -struct Elf64_Rel -{ - /// Address of reference. - Elf64_Addr r_offset; - /// Symbol index and type of relocation. - Elf64_Xword r_info; -} - -/** - * Relocation entry with explicit addend. - */ -struct Elf64_Rela -{ - /// Address of reference. - Elf64_Addr r_offset; - /// Symbol index and type of relocation. - Elf64_Xword r_info; - /// Constant part of expression. - Elf64_Sxword r_addend; -} - -/** - * Symbol table entry. - */ -struct Elf32_Sym -{ - /// Symbol name. - Elf32_Word st_name; - /// Symbol value. - Elf32_Addr st_value; - /// Size of object (e.g., common). - Elf32_Word st_size; - /// Type and Binding attributes. - ubyte st_info; - /// Reserved. - ubyte st_other; - /// Section table index. - Elf32_Half st_shndx; -} - -/** - * Relocation entry. - */ -struct Elf32_Rel -{ - /// Address of reference. - Elf32_Addr r_offset; - /// Symbol index and type of relocation. - Elf32_Word r_info; -} - -/** - * Relocation entry with explicit addend. - */ -struct Elf32_Rela -{ - /// Address of reference. - Elf32_Addr r_offset; - /// Symbol index and type of relocation. - Elf32_Word r_info; - /// Constant part of expression. - Elf32_Sword r_addend; -} - -/// Section Types, sh_type. -enum : Elf64_Word -{ - /// Marks an unused section header. - SHT_NULL = 0, - /// Contains information defined by the program. - SHT_PROGBITS = 1, - /// Contains a linker symbol table. - SHT_SYMTAB = 2, - /// Contains a string table. - SHT_STRTAB = 3, - /// Contains “Rela” type relocation entries. - SHT_RELA = 4, - /// Contains a symbol hash table - SHT_HASH = 5, - /// Contains dynamic linking tables - SHT_DYNAMIC = 6, - /// Contains note information - SHT_NOTE = 7, - /// Contains uninitialized space; does not occupy any space in the file. - SHT_NOBITS = 8, - /// Contains "Rel" type relocation entries. - SHT_REL = 9, - /// Reserved. - SHT_SHLIB = 10, - /// Contains a dynamic loader symbol table. - SHT_DYNSYM = 11, - /// Environment-specific use. - SHT_LOOS = 0x60000000, - SHT_HIOS = 0x6FFFFFFF, - /// Processor-specific use. - SHT_LOPROC = 0x70000000, - SHT_HIPROC = 0x7FFFFFFF, -} - -/** - * Section Attributes, sh_flags. - */ -enum : Elf64_Xword -{ - /// Section contains writable data. - SHF_WRITE = 0x1, - /// Section is allocated in memory image of program. - SHF_ALLOC = 0x2, - /// Section contains executable instructions. - SHF_EXECINSTR = 0x4, - /// Environment-specific use. - SHF_MASKOS = 0x0F000000, - /// Processor-specific use. - SHF_MASKPROC = 0xF0000000, -} - -ubyte ELF64_R_SYM(Elf64_Xword i) @nogc nothrow pure @safe -{ - return cast(ubyte) (i >> 32); -} - -Elf64_Xword ELF64_R_TYPE(Elf64_Xword i) @nogc nothrow pure @safe -{ - return i & 0xffffffffL; -} - -Elf64_Xword ELF64_R_INFO(Elf64_Xword s, Elf64_Xword t) @nogc nothrow pure @safe -{ - return (s << 32) + (t & 0xffffffffL); -} - -ubyte ELF32_ST_BIND(ubyte i) @nogc nothrow pure @safe -{ - return i >> 4; -} - -ubyte ELF32_ST_TYPE(ubyte i) @nogc nothrow pure @safe -{ - return i & 0xf; -} - -ubyte ELF32_ST_INFO(Elf32_Word b, ubyte t) @nogc nothrow pure @safe -{ - return cast(ubyte) ((b << 4) + (t & 0xf)); -} - -Elf32_Word ELF32_R_SYM(Elf32_Word i) @nogc nothrow pure @safe -{ - return i >> 8; -} - -ubyte ELF32_R_TYPE(Elf32_Word i) @nogc nothrow pure @safe -{ - return cast(ubyte) i; -} - -Elf32_Word ELF32_R_INFO(Elf32_Word s, Elf32_Word t) @nogc nothrow pure @safe -{ - return (s << 8) + t; -} - -enum : uint -{ - /// Not visible outside the object file. - STB_LOCAL = 0, - /// Global symbol, visible to all object files. - STB_GLOBAL = 1, - /// Global scope, but with lower precedence than global symbols. - STB_WEAK = 2, - /// Environment-specific use. - STB_LOOS = 10, - STB_HIOS = 12, - /// Processor-specific use. - STB_LOPROC = 13, - STB_HIPROC = 15, -} - -enum : uint -{ - /// No type specified (e.g., an absolute symbol). - STT_NOTYPE = 0, - /// Data object. - STT_OBJECT = 1, - /// Function entry point. - STT_FUNC = 2, - /// Symbol is associated with a section. - STT_SECTION = 3, - /// Source file associated with the object file. - STT_FILE = 4, - /// Environment-specific use. - STT_LOOS = 10, - STT_HIOS = 12, - /// Processor-specific use. - STT_LOPROC = 13, - STT_HIPROC = 15, -} - -/// Special Section Indices. -enum : ushort -{ - /// Used to mark an undefined or meaningless section reference. - SHN_UNDEF = 0, - /// This value specifies the lower bound of the range of reserved indexes. - SHN_LORESERVE = 0xff00, - /// Processor-specific use. - SHN_LOPROC = 0xFF00, - SHN_HIPROC = 0xFF1F, - /// Environment-specific use. - SHN_LOOS = 0xFF20, - SHN_HIOS = 0xFF3F, - /// Indicates that the corresponding reference is an absolute value. - SHN_ABS = 0xFFF1, - /** - * Indicates a symbol that has been declared as a common block (Fortran - * COMMON or C tentative declaration). - */ - SHN_COMMON = 0xFFF2, -} - -/** - * Object File Classes, e_ident[EI_CLASS]. - */ -enum : ubyte -{ - /// Invalid class. - ELFCLASSNONE = 0, - /// 32-bit objects. - ELFCLASS32 = 1, - /// 64-bit objects. - ELFCLASS64 = 2 -} - -enum : ubyte { - /// Invalid version. - EV_NONE = 0, - /// Current version. - EV_CURRENT = 1 -} - -/** - * Data Encodings, e_ident[EI_DATA]. - */ -enum : ubyte -{ - /// Object file data structures are little-endian. - ELFDATA2LSB = 1, - /// Object file data structures are big-endian. - ELFDATA2MSB = 2, -} - -/** - * Operating System and ABI Identifiers, e_ident[EI_OSABI]. - */ -enum EI_OSABI : ubyte -{ - /// System V ABI. - ELFOSABI_SYSV = 0, - /// HP-UX operating system. - ELFOSABI_HPUX = 1, - /// Standalone (embedded) application. - ELFOSABI_STANDALONE = 255, -} - -enum : Elf64_Half -{ - ET_NONE = 0, /// No file type. - ET_REL = 1, /// Relocatable object file. - ET_EXEC = 2, /// Executable file. - ET_DYN = 3, /// Shared object file. - ET_CORE = 4, /// Core file. - ET_LOOS = 0xFE00, /// Environment-specific use. - ET_HIOS = 0xFEFF, - ET_LOPROC = 0xFF00, /// Processor-specific use. - ET_HIPROC = 0xFFFF, -} - -enum : ubyte -{ - R_RISCV_NONE = 0, - /// 32-bit relocation. - R_RISCV_32 = 1, - /// 64-bit relocation. - R_RISCV_64 = 2, - /// Relocation against a local symbol in a shared object. - R_RISCV_RELATIVE = 3, - /// Must be in executable; not allowed in shared library. - R_RISCV_COPY = 4, - /// Indicates the symbol associated with a PLT entry. - R_RISCV_JUMP_SLOT = 5, - R_RISCV_TLS_DTPMOD32 = 6, - R_RISCV_TLS_DTPMOD64 = 7, - R_RISCV_TLS_DTPREL32 = 8, - R_RISCV_TLS_DTPREL64 = 9, - R_RISCV_TLS_TPREL32 = 10, - R_RISCV_TLS_TPREL64 = 11, - /// 12-bit PC-relative branch offset. - R_RISCV_BRANCH = 16, - /// 20-bit PC-relative jump offset. - R_RISCV_JAL = 17, - /// 32-bit PC-relative function call, macros `call`, `tail`. - R_RISCV_CALL = 18, - /// 32-bit PC-relative function call, macros `call`, `tail` (PIC). - R_RISCV_CALL_PLT = 19, - /// High 20 bits of 32-bit PC-relative GOT access, `%got_pcrel_hi(symbol)`. - R_RISCV_GOT_HI20 = 20, - /// High 20 bits of 32-bit PC-relative TLS IE GOT access, macro `la.tls.ie`. - R_RISCV_TLS_GOT_HI20 = 21, - /// High 20 bits of 32-bit PC-relative TLS GD GOT reference, macro `la.tls.gd`. - R_RISCV_TLS_GD_HI20 = 22, - /// High 20 bits of 32-bit PC-relative reference, `%pcrel_hi(symbol)`. - R_RISCV_PCREL_HI20 = 23, - /// Low 12 bits of a 32-bit PC-relative, `%pcrel_lo(address of %pcrel_hi)`, the addend must be 0. - R_RISCV_PCREL_LO12_I = 24, - /// Low 12 bits of a 32-bit PC-relative, `%pcrel_lo(address of %pcrel_hi)`, the addend must be 0. - R_RISCV_PCREL_LO12_S = 25, - /// High 20 bits of 32-bit absolute address, `%hi(symbol)`. - R_RISCV_HI20 = 26, - /// Low 12 bits of 32-bit absolute address, `%lo(symbol)`. - R_RISCV_LO12_I = 27, - /// Low 12 bits of 32-bit absolute address, `%lo(symbol)`. - R_RISCV_LO12_S = 28, - /// High 20 bits of TLS LE thread pointer offset, `%tprel_hi(symbol)`. - R_RISCV_TPREL_HI20 = 29, - /// Low 12 bits of TLS LE thread pointer offset, `%tprel_lo(symbol)`. - R_RISCV_TPREL_LO12_I = 30, - /// Low 12 bits of TLS LE thread pointer offset, `%tprel_lo(symbol)`. - R_RISCV_TPREL_LO12_S = 31, - /// TLS LE thread pointer usage, `%tprel_add(symbol)`. - R_RISCV_TPREL_ADD = 32, - /// 8-bit label addition. - R_RISCV_ADD8 = 33, - /// 16-bit label addition. - R_RISCV_ADD16 = 34, - /// 32-bit label addition. - R_RISCV_ADD32 = 35, - /// 64-bit label addition. - R_RISCV_ADD64 = 36, - /// 8-bit label subtraction. - R_RISCV_SUB8 = 37, - /// 16-bit label subtraction. - R_RISCV_SUB16 = 38, - /// 32-bit label subtraction. - R_RISCV_SUB32 = 39, - /// 64-bit label subtraction. - R_RISCV_SUB64 = 40, - /// GNU {Cpp} vtable hierarchy. - R_RISCV_GNU_VTINHERIT = 41, - /// GNU {Cpp} vtable member usage. - R_RISCV_GNU_VTENTRY = 42, - /// Alignment statement. - R_RISCV_ALIGN = 43, - /// 8-bit PC-relative branch offset. - R_RISCV_RVC_BRANCH = 44, - /// 11-bit PC-relative jump offset. - R_RISCV_RVC_JUMP = 45, - /// High 6 bits of 18-bit absolute address. - R_RISCV_RVC_LUI = 46, - /// Instruction can be relaxed, paired with a normal relocation at the same address. - R_RISCV_RELAX = 51, - /// Local label subtraction. - R_RISCV_SUB6 = 52, - /// Local label assignment. - R_RISCV_SET6 = 53, - /// Local label assignment. - R_RISCV_SET8 = 54, - /// Local label assignment. - R_RISCV_SET16 = 55, - /// Local label assignment. - R_RISCV_SET32 = 56, - /// 32-bit PC relative. - R_RISCV_32_PCREL = 57, - /// Relocation against a local ifunc symbol in a shared object. - R_RISCV_IRELATIVE = 58 -} - -auto pad(ubyte elfClass)(size_t value) @nogc -{ - static if (elfClass == ELFCLASS32) - { - return cast(Elf32_Word) (value / 4 + 1) * 4; - } - else static if (elfClass == ELFCLASS64) - { - return cast(Elf64_Xword) (value / 8 + 1) * 8; - } - else - { - static assert(false, "Invalid ELF class"); - } -} - -private struct Relocation(Sym, Rel) -{ - Sym symbol; - Array!Rel relocations; -} - -struct Elf(ubyte elfClass) -{ - static if (elfClass == ELFCLASS32) - { - alias Addr = Elf32_Addr; - alias Off = Elf32_Off; - alias Half = Elf32_Half; - alias Word = Elf32_Word; - alias Sword = Elf32_Sword; - alias Xword = Elf32_Word; - alias Sxword = Elf32_Sword; - - alias Ehdr = Elf32_Ehdr; - alias Shdr = Elf32_Shdr; - alias Rel = Elf32_Rel; - alias Rela = Elf32_Rela; - alias Sym = Elf32_Sym; - - alias R_SYM = ELF32_R_SYM; - alias R_TYPE = ELF32_R_TYPE; - alias R_INFO = ELF32_R_INFO; - alias ST_BIND = ELF32_ST_BIND; - alias ST_TYPE = ELF32_ST_TYPE; - alias ST_INFO = ELF32_ST_INFO; - } - else static if (elfClass == ELFCLASS64) - { - alias Addr = Elf64_Addr; - alias Off = Elf64_Off; - alias Half = Elf64_Half; - alias Word = Elf64_Word; - alias Sword = Elf64_Sword; - alias Xword = Elf64_Xword; - alias Sxword = Elf64_Sxword; - - alias Ehdr = Elf64_Ehdr; - alias Shdr = Elf64_Shdr; - alias Rel = Elf64_Rel; - alias Rela = Elf64_Rela; - alias Sym = Elf64_Sym; - - alias R_SYM = ELF64_R_SYM; - alias R_TYPE = ELF64_R_TYPE; - alias R_INFO = ELF64_R_INFO; - alias ST_BIND = ELF64_ST_BIND; - alias ST_TYPE = ELF64_ST_TYPE; - alias ST_INFO = ELF64_ST_INFO; - } - else - { - static assert(false, "Invalid ELF class"); - } - - private alias Relocation = .Relocation!(Sym, Rela); - - private Array!Shdr sectionHeaders; - private Off currentOffset = Elf32_Ehdr.sizeof; - static immutable char[52] sections = - "\0.symtab\0.strtab\0.shstrtab\0.text\0.rodata\0.rela.text\0"; - private String strings; - private File output; - private Array!ubyte readOnly; - - private HashTable!(String, Relocation) symbolTable; - - private enum HeaderName - { - text = 0x1b, - roData = 0x21, - string_ = 0x09, - headerString = 0x11, - symbol = 0x01, - rela = 0x29 - } - - static Elf opCall(File output) @nogc - { - Elf elf = Elf.init; - - elf.initializeSectionHeaders(); - elf.output = move(output); - - elf.output.seek(Ehdr.sizeof, File.Whence.set); - - elf.makeTextHeader(); - elf.makeRoDataHeader(); - elf.makeSymbolHeader(); - elf.makeRelaHeader(); - elf.makeStringHeader!(HeaderName.string_)(); - elf.makeStringHeader!(HeaderName.headerString)(); - - return elf; - } - - @disable this(this); - - void finish() @nogc - { - writeRoDataTable(); - writeSymbolTable(); - writeStringTables(); - - // End writing data, start writing headers. - - output.write((cast(ubyte*) this.sectionHeaders.get)[0 .. Shdr.sizeof * this.sectionHeaders.length]); - - writeFileHeader(); - } - - private Sym initializeSymbols() @nogc - { - // Zero symbol - Sym symbol; - symbol.st_name = 0; // Word - symbol.st_value = 0; // Addr - symbol.st_size = 0; // Word - symbol.st_info = 0; // char - symbol.st_other = 0; // char - symbol.st_shndx = 0; // Half word - - return symbol; - } - - private void makeStringHeader(HeaderName position)() @nogc - { - Shdr table; - - table.sh_name = position; - table.sh_type = SHT_STRTAB; - table.sh_flags = 0; - table.sh_addr = 0; - table.sh_offset = 0; - table.sh_size = 0; - table.sh_link = SHN_UNDEF; - table.sh_info = 0; - table.sh_addralign = 1; - table.sh_entsize = 0; - - this.sectionHeaders.insertBack(table); - } - - private void writeStringTables() @nogc - { - auto stringIndex = findHeader!(HeaderName.string_); - assert(stringIndex != -1); - - this.sectionHeaders[stringIndex].sh_offset = this.currentOffset; - this.sectionHeaders[stringIndex].sh_size = cast(Word) strings.length; - - output.write(cast(ubyte[]) this.strings.toStringz[0 .. this.strings.length + 1]); - this.currentOffset += this.strings.length + 1; - - auto headerStringIndex = findHeader!(HeaderName.headerString); - assert(stringIndex != -1); - - this.sectionHeaders[headerStringIndex].sh_offset = this.currentOffset; - this.sectionHeaders[headerStringIndex].sh_size = cast(Word) sections.length; - - output.write(cast(const(ubyte)[]) this.sections); - this.currentOffset += this.sections.length; - auto alignment = pad!ELFCLASS32(this.strings.length + 1 + this.sections.length); - const(ubyte)[4] padding = 0; - output.write(padding[0 .. alignment - this.strings.length - 1 - this.sections.length]); - this.currentOffset += alignment - this.strings.length - 1 - this.sections.length; - } - - private void makeSymbolHeader() @nogc - { - Shdr symbolTableHeader; - - symbolTableHeader.sh_name = HeaderName.symbol; - symbolTableHeader.sh_type = SHT_SYMTAB; - symbolTableHeader.sh_flags = 0; - symbolTableHeader.sh_addr = 0; - symbolTableHeader.sh_offset = 0; - symbolTableHeader.sh_size = 0; - // String table used by entries in this section. - symbolTableHeader.sh_link = 0; - symbolTableHeader.sh_info = 0; - symbolTableHeader.sh_addralign = 4; - symbolTableHeader.sh_entsize = Sym.sizeof; - - this.sectionHeaders.insertBack(symbolTableHeader); - } - - private void writeSymbolTable() @nogc - { - const index = findHeader!(HeaderName.symbol)(); - const stringIndex = findHeader!(HeaderName.string_)(); - const relaIndex = findHeader!(HeaderName.rela); - const textIndex = findHeader!(HeaderName.text)(); - - assert(index != -1); - assert(stringIndex != -1); - assert(relaIndex != -1); - assert(textIndex != -1); - - this.sectionHeaders[index].sh_offset = this.currentOffset; - this.sectionHeaders[index].sh_link = cast(Word) stringIndex; - this.sectionHeaders[index].sh_size = cast(Word) ((1 + symbolTable.length) * Sym.sizeof); - - this.sectionHeaders[relaIndex].sh_link = cast(Word) index; - this.sectionHeaders[relaIndex].sh_info = cast(Word) textIndex; - this.sectionHeaders[relaIndex].sh_offset = this.sectionHeaders[index].sh_offset - + this.sectionHeaders[index].sh_size; - - auto initialSymbol = initializeSymbols(); - output.write((cast(ubyte*) &initialSymbol)[0 .. Sym.sizeof]); - this.currentOffset += Sym.sizeof; - - int i = 1; - Array!Relocation symbols = Array!Relocation(this.symbolTable.byValue()); - auto rightRange = symbols[].partition!(symbol => ST_BIND(symbol.symbol.st_info) != STB_GLOBAL); - - // Greater than last local symbol. - this.sectionHeaders[index].sh_info = cast(Word) (symbols.length - rightRange.length + 1); - - foreach (ref symbol; symbols[]) - { - this.output.seek(this.sectionHeaders[relaIndex].sh_offset + this.sectionHeaders[relaIndex].sh_size, - File.Whence.set); - - if (!symbol.relocations.empty) - { - foreach (ref relocation; symbol.relocations[]) - { - relocation.r_info = R_INFO(i, R_TYPE(relocation.r_info)); - } - this.sectionHeaders[relaIndex].sh_flags = SHF_ALLOC; - const size = cast(Word) (Rela.sizeof * symbol.relocations.length); - - this.output.write((cast(ubyte*) symbol.relocations.get)[0 .. size]); - this.sectionHeaders[relaIndex].sh_size += size; - this.currentOffset += size; - } - - this.output.seek(this.sectionHeaders[index].sh_offset + i * Sym.sizeof, File.Whence.set); - output.write((cast(ubyte*) &symbol)[0 .. Sym.sizeof]); - this.currentOffset += Sym.sizeof; - ++i; - } - this.output.seek(0, File.Whence.end); - } - - void addCode(String name, ref Array!ubyte text) - @nogc - { - this.output.write(text.get); - - auto textHeaderIndex = findHeader!(HeaderName.text)(); - assert(textHeaderIndex != -1); - - this.strings.insertBack("\0"); - - Sym symbol; - // Main function - symbol.st_name = cast(Word) this.strings.length; - symbol.st_value = 0; - symbol.st_size = cast(Word) text.length; - symbol.st_info = ST_INFO(STB_GLOBAL, STT_FUNC); - symbol.st_other = 0; // char - // .text header index, half word - symbol.st_shndx = cast(Half) textHeaderIndex; - this.symbolTable[name] = Relocation(symbol); - - this.strings.insertBack(name[]); - - this.sectionHeaders[textHeaderIndex].sh_size += text.length; - this.currentOffset += text.length; - } - - void addReadOnlyData(String name, ref Array!ubyte data) @nogc - { - auto roDataIndex = findHeader!(HeaderName.roData)(); - assert(roDataIndex != -1); - - this.strings.insertBack("\0"); - - Sym symbol; - // Main function - symbol.st_name = cast(Word) this.strings.length; - symbol.st_value = 0; - symbol.st_size = cast(Word) data.length; - symbol.st_info = ST_INFO(STB_LOCAL, STT_NOTYPE); - symbol.st_other = 0; // char - // .text header index, half word - symbol.st_shndx = cast(Half) roDataIndex; - this.symbolTable[name] = Relocation(symbol); - - this.strings.insertBack(name[]); - this.readOnly.insertBack(data[]); - } - - void addExternSymbol(String name) @nogc - { - Sym usedSymbolEntry; - - this.strings.insertBack("\0"); - usedSymbolEntry.st_name = cast(Word) this.strings.length; - usedSymbolEntry.st_value = 0; - usedSymbolEntry.st_size = 0; - usedSymbolEntry.st_info = ST_INFO(STB_GLOBAL, STT_NOTYPE); - usedSymbolEntry.st_other = 0; - usedSymbolEntry.st_shndx = SHN_UNDEF; - - this.strings.insertBack(name[]); - this.strings.insertBack("\0"); - this.symbolTable[name] = Relocation(usedSymbolEntry); - } - - void relocate(String name, Rela[] usedSymbols...) @nogc - { - foreach (usedSymbol; usedSymbols) - { - Rela relocationEntry = usedSymbol; - - relocationEntry.r_info = usedSymbol.r_info; - this.symbolTable[name].relocations.insertBack(relocationEntry); - } - } - - private ptrdiff_t findHeader(HeaderName position)() - { - return countUntil!(header => header.sh_name == position)(this.sectionHeaders[]); - } - - private void makeTextHeader() @nogc - { - Shdr textHeader; - - textHeader.sh_name = HeaderName.text; - textHeader.sh_type = SHT_PROGBITS; - textHeader.sh_flags = SHF_EXECINSTR | SHF_ALLOC; - textHeader.sh_addr = 0; - textHeader.sh_offset = this.currentOffset; - textHeader.sh_size = 0; - textHeader.sh_link = SHN_UNDEF; - textHeader.sh_info = 0; - textHeader.sh_addralign = 1; - textHeader.sh_entsize = 0; - - this.sectionHeaders.insertBack(textHeader); - } - - private void initializeSectionHeaders() @nogc - { - Shdr table; - - table.sh_name = 0; - table.sh_type = SHT_NULL; - table.sh_flags = 0; - table.sh_addr = 0; - table.sh_offset = 0; - table.sh_size = 0; - table.sh_link = SHN_UNDEF; - table.sh_info = 0; - table.sh_addralign = 0; - table.sh_entsize = 0; - - this.sectionHeaders.insertBack(table); - } - - private void writeFileHeader() @nogc - { - Ehdr fileHeader; - auto headerStringIndex = findHeader!(HeaderName.headerString)(); - - assert(headerStringIndex != -1); - - // Magic number. - fileHeader.e_ident[0] = '\x7f'; - fileHeader.e_ident[1] = 'E'; - fileHeader.e_ident[2] = 'L'; - fileHeader.e_ident[3] = 'F'; - - fileHeader.e_ident[4] = ELFCLASS32; - fileHeader.e_ident[5] = ELFDATA2LSB; - fileHeader.e_ident[6] = EV_CURRENT; - fileHeader.e_ident[7] = EI_OSABI.ELFOSABI_SYSV; - fileHeader.e_ident[8] = 0; - - fileHeader.e_type = ET_REL; - fileHeader.e_machine = 0xf3; // EM_RISCV - fileHeader.e_version = EV_CURRENT; - fileHeader.e_entry = 0; - fileHeader.e_phoff = 0; - fileHeader.e_shoff = this.currentOffset; - fileHeader.e_flags = 0; - fileHeader.e_ehsize = Elf32_Ehdr.sizeof; - fileHeader.e_phentsize = 0; - fileHeader.e_phnum = 0; - fileHeader.e_shentsize = Elf32_Shdr.sizeof; - fileHeader.e_shnum = cast(Elf32_Half) this.sectionHeaders.length; - - // String table is the last one - fileHeader.e_shstrndx = cast(Half) headerStringIndex; - - output.seek(0, File.Whence.set); - output.write((cast(ubyte*) &fileHeader)[0 .. fileHeader.sizeof]); - } - - private void makeRoDataHeader() @nogc - { - Shdr table; - - table.sh_name = HeaderName.roData; - table.sh_type = SHT_PROGBITS; - table.sh_flags = SHF_ALLOC; - table.sh_addr = 0; - table.sh_offset = 0; - table.sh_size = 0; - table.sh_link = SHN_UNDEF; - table.sh_info = 0; - table.sh_addralign = 4; - table.sh_entsize = 0; - - this.sectionHeaders.insertBack(table); - } - - private void writeRoDataTable() @nogc - { - auto index = findHeader!(HeaderName.roData)(); - assert(index != -1); - - this.sectionHeaders[index].sh_offset = this.currentOffset; - this.sectionHeaders[index].sh_size = cast(Xword) this.readOnly.length; - - output.write(this.readOnly.get); - this.currentOffset += this.readOnly.length; - } - - private void makeRelaHeader() @nogc - { - Shdr table; - - table.sh_name = HeaderName.rela; - table.sh_type = SHT_RELA; - table.sh_flags = 0; - table.sh_addr = 0; - table.sh_offset = 0; - table.sh_size = 0; - table.sh_link = SHN_UNDEF; - table.sh_info = 0; - table.sh_addralign = 4; - table.sh_entsize = Rela.sizeof; - - this.sectionHeaders.insertBack(table); - } -} diff --git a/source/elna/extended.d b/source/elna/extended.d deleted file mode 100644 index ff06a0c..0000000 --- a/source/elna/extended.d +++ /dev/null @@ -1,335 +0,0 @@ -/** - * File I/O that can be moved into more generic library when and if finished. - */ -module elna.extended; - -import core.stdc.errno; -import core.stdc.stdio; -import std.sumtype; -import std.typecons; -import tanya.os.error; -import tanya.container.array; -import tanya.container.string; - -/** - * File handle abstraction. - */ -struct File -{ - /// Plattform dependent file type. - alias Handle = FILE*; - - /// Uninitialized file handle value. - enum Handle invalid = null; - - /** - * Relative position. - */ - enum Whence - { - /// Relative to the start of the file. - set = SEEK_SET, - /// Relative to the current cursor position. - currentt = SEEK_CUR, - /// Relative from the end of the file. - end = SEEK_END, - } - - /** - * File open modes. - */ - enum Mode - { - /// Open the file for reading. - read = 1 << 0, - /// Open the file for writing. The stream is positioned at the beginning - /// of the file. - write = 1 << 1, - /// Open the file for writing and remove its contents. - truncate = 1 << 2, - /// Open the file for writing. The stream is positioned at the end of - /// the file. - append = 1 << 3, - } - - private enum Status - { - invalid, - owned, - borrowed, - } - - private union Storage - { - Handle handle; - ErrorCode errorCode; - } - private Storage storage; - private Status status = Status.invalid; - - @disable this(scope return ref File f); - @disable this(); - - /** - * Closes the file. - */ - ~this() @nogc nothrow - { - if (this.status == Status.owned) - { - fclose(this.storage.handle); - } - this.storage.handle = invalid; - this.status = Status.invalid; - } - - /** - * Construct the object with the given system handle. The won't be claused - * in the descructor if this constructor is used. - * - * Params: - * handle = File handle to be wrapped by this structure. - */ - this(Handle handle) @nogc nothrow pure @safe - { - this.storage.handle = handle; - this.status = Status.borrowed; - } - - /** - * Returns: Plattform dependent file handle. - */ - @property Handle handle() @nogc nothrow pure @trusted - { - return valid ? this.storage.handle : invalid; - } - - /** - * Returns: An error code if an error has occurred. - */ - @property ErrorCode errorCode() @nogc nothrow pure @safe - { - return valid ? ErrorCode() : this.storage.errorCode; - } - - /** - * Returns: Whether a valid, opened file is represented. - */ - @property bool valid() @nogc nothrow pure @safe - { - return this.status != Status.invalid; - } - - /** - * Transfers the file into invalid state. - * - * Returns: The old file handle. - */ - Handle reset() @nogc nothrow pure @safe - { - if (!valid) - { - return invalid; - } - auto oldHandle = handle; - - this.status = Status.invalid; - this.storage.errorCode = ErrorCode(); - - return oldHandle; - } - - /** - * Sets stream position in the file. - * - * Params: - * offset = File offset. - * whence = File position to add the offset to. - * - * Returns: Error code if any. - */ - ErrorCode seek(size_t offset, Whence whence) @nogc nothrow - { - if (!valid) - { - return ErrorCode(ErrorCode.ErrorNo.badDescriptor); - } - if (fseek(this.storage.handle, offset, whence)) - { - return ErrorCode(cast(ErrorCode.ErrorNo) errno); - } - return ErrorCode(); - } - - /** - * Returns: Current offset or an error. - */ - SumType!(ErrorCode, size_t) tell() @nogc nothrow - { - if (!valid) - { - return typeof(return)(ErrorCode(ErrorCode.ErrorNo.badDescriptor)); - } - auto result = ftell(this.storage.handle); - - if (result < 0) - { - return typeof(return)(ErrorCode(cast(ErrorCode.ErrorNo) errno)); - } - return typeof(return)(cast(size_t) result); - } - - /** - * Params: - * buffer = Destination buffer. - * - * Returns: Bytes read. $(D_PSYMBOL ErrorCode.ErrorNo.success) means that - * while reading the file an unknown error has occurred. - */ - SumType!(ErrorCode, size_t) read(ubyte[] buffer) @nogc nothrow - { - if (!valid) - { - return typeof(return)(ErrorCode(ErrorCode.ErrorNo.badDescriptor)); - } - const bytesRead = fread(buffer.ptr, 1, buffer.length, this.storage.handle); - if (bytesRead == buffer.length || eof()) - { - return typeof(return)(bytesRead); - } - return typeof(return)(ErrorCode()); - } - - /** - * Params: - * buffer = Source buffer. - * - * Returns: Bytes written. $(D_PSYMBOL ErrorCode.ErrorNo.success) means that - * while reading the file an unknown error has occurred. - */ - SumType!(ErrorCode, size_t) write(const(ubyte)[] buffer) @nogc nothrow - { - if (!valid) - { - return typeof(return)(ErrorCode(ErrorCode.ErrorNo.badDescriptor)); - } - const bytesWritten = fwrite(buffer.ptr, 1, buffer.length, this.storage.handle); - if (bytesWritten == buffer.length) - { - return typeof(return)(buffer.length); - } - return typeof(return)(ErrorCode()); - } - - /** - * Returns: EOF status of the file. - */ - bool eof() @nogc nothrow - { - return valid && feof(this.storage.handle) != 0; - } - - /** - * Constructs a file object that will be closed in the destructor. - * - * Params: - * filename = The file to open. - * - * Returns: Opened file or an error. - */ - static File open(const(char)* filename, BitFlags!Mode mode) @nogc nothrow - { - char[3] modeBuffer = "\0\0\0"; - - if (mode.truncate) - { - modeBuffer[0] = 'w'; - if (mode.read) - { - modeBuffer[1] = '+'; - } - } - else if (mode.append) - { - modeBuffer[0] = 'a'; - if (mode.read) - { - modeBuffer[1] = '+'; - } - } - else if (mode.read) - { - modeBuffer[0] = 'r'; - if (mode.write) - { - modeBuffer[1] = '+'; - } - } - - auto newHandle = fopen(filename, modeBuffer.ptr); - auto newFile = File(newHandle); - - if (newHandle is null) - { - newFile.status = Status.invalid; - newFile.storage.errorCode = ErrorCode(cast(ErrorCode.ErrorNo) errno); - } - else - { - if (mode == BitFlags!Mode(Mode.write)) - { - rewind(newHandle); - } - newFile.status = Status.owned; - } - - return newFile; - } -} - -/** - * Reads the whole file and returns its contents. - * - * Params: - * sourceFilename = Source filename. - * - * Returns: File contents or an error. - * - * See_Also: $(D_PSYMBOL File.read) - */ -SumType!(ErrorCode, Array!ubyte) readFile(String sourceFilename) @nogc -{ - enum size_t bufferSize = 255; - auto sourceFile = File.open(sourceFilename.toStringz, BitFlags!(File.Mode)(File.Mode.read)); - - if (!sourceFile.valid) - { - return typeof(return)(sourceFile.errorCode); - } - Array!ubyte sourceText; - size_t totalRead; - size_t bytesRead; - do - { - sourceText.length = sourceText.length + bufferSize; - const readStatus = sourceFile - .read(sourceText[totalRead .. $].get) - .match!( - (ErrorCode errorCode) => nullable(errorCode), - (size_t bytesRead_) { - bytesRead = bytesRead_; - return Nullable!ErrorCode(); - } - ); - if (!readStatus.isNull) - { - return typeof(return)(readStatus.get); - } - totalRead += bytesRead; - } - while (bytesRead == bufferSize); - - sourceText.length = totalRead; - - return typeof(return)(sourceText); -} diff --git a/source/elna/ir.d b/source/elna/ir.d deleted file mode 100644 index edb70fa..0000000 --- a/source/elna/ir.d +++ /dev/null @@ -1,272 +0,0 @@ -module elna.ir; - -import parser = elna.parser; -import tanya.container.array; -import tanya.container.hashtable; -import tanya.container.string; -import tanya.memory.allocator; -public import elna.parser : BinaryOperator; - -/** - * Mapping between the parser and IR AST. - */ -struct ASTMapping -{ - alias Node = .Node; - alias Definition = .Definition; - alias VariableDeclaration = .VariableDeclaration; - alias Statement = Array!(.Statement); - alias BangStatement = .Expression; - alias Block = .Definition; - alias Expression = .Expression; - alias Number = .Number; - alias Variable = .Variable; - alias BinaryExpression = .BinaryExpression; - -} - -/** - * IR visitor. -*/ -interface IRVisitor -{ - abstract void visit(Node) @nogc; - abstract void visit(Definition) @nogc; - abstract void visit(Expression) @nogc; - abstract void visit(Statement) @nogc; - abstract void visit(Variable) @nogc; - abstract void visit(VariableDeclaration) @nogc; - abstract void visit(Number) @nogc; - abstract void visit(BinaryExpression) @nogc; -} - -/** - * AST node. - */ -abstract class Node -{ - abstract void accept(IRVisitor) @nogc; -} - -/** - * Definition. - */ -class Definition : Node -{ - char[] identifier; - Array!Statement statements; - Array!VariableDeclaration variableDeclarations; - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -class Statement : Node -{ - BinaryExpression expression; - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -abstract class Expression : Node -{ - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -class Number : Expression -{ - int value; - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -class Variable : Expression -{ - size_t counter; - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -class VariableDeclaration : Node -{ - String identifier; - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -class BinaryExpression : Node -{ - Expression lhs, rhs; - BinaryOperator operator; - - this(Expression lhs, Expression rhs, BinaryOperator operator) - @nogc - { - this.lhs = lhs; - this.rhs = rhs; - this.operator = operator; - } - - override void accept(IRVisitor visitor) @nogc - { - visitor.visit(this); - } -} - -final class TransformVisitor : parser.ParserVisitor!ASTMapping -{ - private HashTable!(String, int) constants; - - ASTMapping.Node visit(parser.Node node) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.Definition visit(parser.Definition definition) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.VariableDeclaration visit(parser.VariableDeclaration declaration) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.BangStatement visit(parser.BangStatement statement) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.Block visit(parser.Block block) @nogc - { - auto target = defaultAllocator.make!Definition; - this.constants = transformConstants(block.definitions); - - target.statements = block.statement.accept(this); - target.variableDeclarations = transformVariableDeclarations(block.variableDeclarations); - - return target; - } - - ASTMapping.Expression visit(parser.Expression expression) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.Number visit(parser.Number number) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.Variable visit(parser.Variable variable) @nogc - { - assert(false, "Not implemented"); - } - - ASTMapping.BinaryExpression visit(parser.BinaryExpression) @nogc - { - assert(false, "Not implemented"); - } - - private Number transformNumber(parser.Number number) @nogc - { - return defaultAllocator.make!Number(number.value); - } - - private Variable binaryExpression(parser.BinaryExpression binaryExpression, - ref Array!Statement statements) @nogc - { - auto target = defaultAllocator.make!BinaryExpression( - expression(binaryExpression.lhs, statements), - expression(binaryExpression.rhs, statements), - binaryExpression.operator - ); - - auto newStatement = defaultAllocator.make!Statement; - newStatement.expression = target; - statements.insertBack(newStatement); - - auto newVariable = defaultAllocator.make!Variable; - newVariable.counter = statements.length; - - return newVariable; - } - - private Expression expression(parser.Expression expression, - ref Array!Statement statements) @nogc - { - if ((cast(parser.Number) expression) !is null) - { - auto numberExpression = defaultAllocator.make!Number; - numberExpression.value = (cast(parser.Number) expression).value; - - return numberExpression; - } - if ((cast(parser.Variable) expression) !is null) - { - auto numberExpression = defaultAllocator.make!Number; - numberExpression.value = this.constants[(cast(parser.Variable) expression).identifier]; - - return numberExpression; - } - else if ((cast(parser.BinaryExpression) expression) !is null) - { - return binaryExpression(cast(parser.BinaryExpression) expression, statements); - } - return null; - } - - override Array!Statement visit(parser.Statement statement) @nogc - { - typeof(return) statements; - if ((cast(parser.BangStatement) statement) !is null) - { - expression((cast(parser.BangStatement) statement).expression, statements); - } - return statements; - } - - private HashTable!(String, int) transformConstants(ref Array!(parser.Definition) definitions) @nogc - { - typeof(return) constants; - - foreach (definition; definitions[]) - { - constants[definition.identifier] = definition.number.value; - } - - return constants; - } - - Array!VariableDeclaration transformVariableDeclarations(ref Array!(parser.VariableDeclaration) variableDeclarations) - @nogc - { - typeof(return) variables; - - foreach (ref variableDeclaration; variableDeclarations) - { - auto newDeclaration = defaultAllocator.make!VariableDeclaration; - newDeclaration.identifier = variableDeclaration.identifier; - variables.insertBack(newDeclaration); - } - - return variables; - } -} diff --git a/source/elna/lexer.d b/source/elna/lexer.d deleted file mode 100644 index 83fb873..0000000 --- a/source/elna/lexer.d +++ /dev/null @@ -1,254 +0,0 @@ -module elna.lexer; - -import core.stdc.stdlib; -import core.stdc.ctype; -import core.stdc.string; -import elna.result; -import std.range; -import tanya.container.array; -import tanya.container.string; - -struct Token -{ - enum Type - { - number, - operator, - let, - identifier, - equals, - var, - semicolon, - leftParen, - rightParen, - bang, - dot, - comma, - } - - union Value - { - int number; - String identifier; - } - - private Type type; - private Value value_; - private Position position_; - - @disable this(); - - this(Type type, Position position) @nogc nothrow pure @safe - { - this.type = type; - this.position_ = position; - } - - this(Type type, int value, Position position) @nogc nothrow pure @trusted - in (type == Type.number) - { - this(type, position); - this.value_.number = value; - } - - this()(Type type, auto ref String value, Position position) - @nogc nothrow pure @trusted - in (type == Type.identifier || type == Type.operator) - { - this(type, position); - this.value_.identifier = value; - } - - /** - * Params: - * type = Expected type. - * - * Returns: Whether this token is of the expected type. - */ - bool ofType(Type type) const @nogc nothrow pure @safe - { - return this.type == type; - } - - @property auto value(Type type)() @nogc nothrow pure @trusted - in (ofType(type)) - { - static if (type == Type.number) - { - return this.value_.number; - } - else static if (type == Type.identifier || type == Type.operator) - { - return this.value_.identifier; - } - else - { - static assert(false, "This type doesn't have a value"); - } - } - - /** - * Returns: The token position in the source text. - */ - @property const(Position) position() const @nogc nothrow pure @safe - { - return this.position_; - } -} - -/** - * Range over the source text that keeps track of the current position. - */ -struct Source -{ - char[] buffer; - Position position; - - this(char[] buffer) @nogc nothrow pure @safe - { - this.buffer = buffer; - } - - @disable this(); - - bool empty() @nogc nothrow pure @safe - { - return this.length == 0; - } - - char front() @nogc nothrow pure @safe - in (!empty) - { - return this.buffer[0]; - } - - void popFront() @nogc nothrow pure @safe - in (!empty) - { - this.buffer = buffer[1 .. $]; - ++this.position.column; - } - - void breakLine() @nogc nothrow pure @safe - in (!empty) - { - this.buffer = buffer[1 .. $]; - ++this.position.line; - this.position.column = 1; - } - - @property size_t length() const @nogc nothrow pure @safe - { - return this.buffer.length; - } - - char opIndex(size_t index) @nogc nothrow pure @safe - in (index < length) - { - return this.buffer[index]; - } - - char[] opSlice(size_t i, size_t j) @nogc nothrow pure @safe - in - { - assert(i <= j); - assert(j <= length); - } - do - { - return this.buffer[i .. j]; - } -} - -Result!(Array!Token) lex(char[] buffer) @nogc -{ - Array!Token tokens; - auto source = Source(buffer); - - while (!source.empty) - { - if (source.front == ' ') - { - source.popFront; - } - else if (source.front >= '0' && source.front <= '9') // Multi-digit. - { - tokens.insertBack(Token(Token.Type.number, source.front - '0', source.position)); - source.popFront; - } - else if (source.front == '=') - { - tokens.insertBack(Token(Token.Type.equals, source.position)); - source.popFront; - } - else if (source.front == '(') - { - tokens.insertBack(Token(Token.Type.leftParen, source.position)); - source.popFront; - } - else if (source.front == ')') - { - tokens.insertBack(Token(Token.Type.rightParen, source.position)); - source.popFront; - } - else if (source.front == ';') - { - tokens.insertBack(Token(Token.Type.semicolon, source.position)); - source.popFront; - } - else if (source.front == ',') - { - tokens.insertBack(Token(Token.Type.comma, source.position)); - source.popFront; - } - else if (source.front == '!') - { - tokens.insertBack(Token(Token.Type.bang, source.position)); - source.popFront; - } - else if (source.front == '.') - { - tokens.insertBack(Token(Token.Type.dot, source.position)); - source.popFront; - } - else if (isalpha(source.front)) - { - size_t i = 1; - while (i < source.length && isalpha(source[i])) - { - ++i; - } - if (source[0 .. i] == "const") - { - tokens.insertBack(Token(Token.Type.let, source.position)); - } - else if (source[0 .. i] == "var") - { - tokens.insertBack(Token(Token.Type.var, source.position)); - } - else - { - auto identifier = String(source[0 .. i]); - tokens.insertBack(Token(Token.Type.identifier, identifier, source.position)); - } - source.popFrontN(i); - } - else if (source.front == '+' || source.front == '-') - { - String operator; - - operator.insertBack(source.front); - tokens.insertBack(Token(Token.Type.operator, operator, source.position)); - source.popFront; - } - else if (source.front == '\n') - { - source.breakLine; - } - else - { - return typeof(return)("Unexptected next character", source.position); - } - } - return typeof(return)(tokens); -} diff --git a/source/elna/parser.d b/source/elna/parser.d deleted file mode 100644 index c0bf048..0000000 --- a/source/elna/parser.d +++ /dev/null @@ -1,372 +0,0 @@ -module elna.parser; - -import elna.lexer; -import elna.result; -import tanya.container.array; -import tanya.container.string; -import tanya.memory.allocator; - - -/** - * Parser visitor. - */ -interface ParserVisitor(Mapping) -{ - Mapping.Node visit(Node) @nogc; - Mapping.Definition visit(Definition) @nogc; - Mapping.VariableDeclaration visit(VariableDeclaration) @nogc; - Mapping.Statement visit(Statement) @nogc; - Mapping.BangStatement visit(BangStatement) @nogc; - Mapping.Block visit(Block) @nogc; - Mapping.Expression visit(Expression) @nogc; - Mapping.Number visit(Number) @nogc; - Mapping.Variable visit(Variable) @nogc; - Mapping.BinaryExpression visit(BinaryExpression) @nogc; -} - -/** - * AST node. - */ -abstract class Node -{ - Mapping.Node accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -/** - * Constant definition. - */ -class Definition : Node -{ - Number number; - String identifier; - - Mapping.Definition accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -/** - * Variable declaration. - */ -class VariableDeclaration : Node -{ - String identifier; - - Mapping.VariableDeclaration accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -abstract class Statement : Node -{ - Mapping.Statement accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -class BangStatement : Statement -{ - Expression expression; - - Mapping.BangStatement accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -class Block : Node -{ - Array!Definition definitions; - Array!VariableDeclaration variableDeclarations; - Statement statement; - - Mapping.Block accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -abstract class Expression : Node -{ - Mapping.Expression accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -class Number : Expression -{ - int value; - - Mapping.Number accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -class Variable : Expression -{ - String identifier; - - Mapping.Variable accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -enum BinaryOperator -{ - sum, - subtraction -} - -class BinaryExpression : Expression -{ - Expression lhs, rhs; - BinaryOperator operator; - - this(Expression lhs, Expression rhs, String operator) @nogc - { - this.lhs = lhs; - this.rhs = rhs; - if (operator == "+") - { - this.operator = BinaryOperator.sum; - } - else if (operator == "-") - { - this.operator = BinaryOperator.subtraction; - } - else - { - assert(false, "Invalid binary operator"); - } - } - - Mapping.BinaryExpression accept(Mapping)(ParserVisitor!Mapping visitor) @nogc - { - return visitor.visit(this); - } -} - -private Result!Expression parseFactor(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected factor, got end of stream") -{ - if (tokens.front.ofType(Token.Type.identifier)) - { - auto variable = defaultAllocator.make!Variable; - variable.identifier = tokens.front.value!(Token.Type.identifier); - tokens.popFront; - return Result!Expression(variable); - } - else if (tokens.front.ofType(Token.Type.number)) - { - auto number = defaultAllocator.make!Number; - number.value = tokens.front.value!(Token.Type.number); - tokens.popFront; - return Result!Expression(number); - } - else if (tokens.front.ofType(Token.Type.leftParen)) - { - tokens.popFront; - - auto expression = parseExpression(tokens); - - tokens.popFront; - return expression; - } - return Result!Expression("Expected a factor", tokens.front.position); -} - -private Result!Expression parseTerm(ref Array!(Token).Range tokens) @nogc -{ - return parseFactor(tokens); -} - -private Result!Expression parseExpression(ref Array!(Token).Range tokens) @nogc -in (!tokens.empty, "Expected expression, got end of stream") -{ - auto term = parseTerm(tokens); - if (!term.valid || tokens.empty || !tokens.front.ofType(Token.Type.operator)) - { - return term; - } - auto operator = tokens.front.value!(Token.Type.operator); - tokens.popFront; - - auto expression = parseExpression(tokens); - - if (expression.valid) - { - auto binaryExpression = defaultAllocator - .make!BinaryExpression(term.result, expression.result, operator); - - return Result!Expression(binaryExpression); - } - else - { - return Result!Expression("Expected right-hand side to be an expression", tokens.front.position); - } -} - -private Result!Definition parseDefinition(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected definition, got end of stream") -{ - auto definition = defaultAllocator.make!Definition; - definition.identifier = tokens.front.value!(Token.Type.identifier); // Copy. - - tokens.popFront(); - tokens.popFront(); // Skip the equals sign. - - if (tokens.front.ofType(Token.Type.number)) - { - auto number = defaultAllocator.make!Number; - number.value = tokens.front.value!(Token.Type.number); - definition.number = number; - tokens.popFront; - return Result!Definition(definition); - } - return Result!Definition("Expected a number", tokens.front.position); -} - -private Result!Statement parseStatement(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected block, got end of stream") -{ - if (tokens.front.ofType(Token.Type.bang)) - { - tokens.popFront; - auto statement = defaultAllocator.make!BangStatement; - auto expression = parseExpression(tokens); - if (expression.valid) - { - statement.expression = expression.result; - } - else - { - return Result!Statement(expression.error.get); - } - return Result!Statement(statement); - } - return Result!Statement("Expected ! statement", tokens.front.position); -} - -private Result!(Array!Definition) parseDefinitions(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected definition, got end of stream") -{ - tokens.popFront; // Skip const. - - Array!Definition definitions; - - while (!tokens.empty) - { - auto definition = parseDefinition(tokens); - if (!definition.valid) - { - return typeof(return)(definition.error.get); - } - definitions.insertBack(definition.result); - if (tokens.front.ofType(Token.Type.semicolon)) - { - break; - } - if (tokens.front.ofType(Token.Type.comma)) - { - tokens.popFront; - } - } - - return typeof(return)(definitions); -} - -private Result!(Array!VariableDeclaration) parseVariableDeclarations(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected variable declarations, got end of stream") -{ - tokens.popFront; // Skip var. - - Array!VariableDeclaration variableDeclarations; - - while (!tokens.empty) - { - auto currentToken = tokens.front; - if (currentToken.ofType(Token.Type.identifier)) - { - auto variableDeclaration = defaultAllocator.make!VariableDeclaration; - variableDeclaration.identifier = currentToken.value!(Token.Type.identifier); - variableDeclarations.insertBack(variableDeclaration); - tokens.popFront; - } - else - { - return typeof(return)("Expected variable name", tokens.front.position); - } - if (tokens.empty) - { - return typeof(return)("Expected \";\" or \",\" name", currentToken.position); - } - if (tokens.front.ofType(Token.Type.semicolon)) - { - break; - } - if (tokens.front.ofType(Token.Type.comma)) - { - tokens.popFront; - } - } - - return typeof(return)(variableDeclarations); -} - -private Result!Block parseBlock(ref Array!Token.Range tokens) @nogc -in (!tokens.empty, "Expected block, got end of stream") -{ - auto block = defaultAllocator.make!Block; - if (tokens.front.ofType(Token.Type.let)) - { - auto constDefinitions = parseDefinitions(tokens); - if (constDefinitions.valid) - { - block.definitions = constDefinitions.result; - } - else - { - return Result!Block(constDefinitions.error.get); - } - tokens.popFront; - } - if (tokens.front.ofType(Token.Type.var)) - { - auto variableDeclarations = parseVariableDeclarations(tokens); - if (variableDeclarations.valid) - { - block.variableDeclarations = variableDeclarations.result; - } - else - { - return Result!Block(variableDeclarations.error.get); - } - tokens.popFront; - } - auto statement = parseStatement(tokens); - if (statement.valid) - { - block.statement = statement.result; - } - else - { - return Result!Block(statement.error.get); - } - - return Result!Block(block); -} - -Result!Block parse(ref Array!Token tokenStream) @nogc -{ - auto tokens = tokenStream[]; - return parseBlock(tokens); -} diff --git a/source/elna/result.d b/source/elna/result.d deleted file mode 100644 index 9427147..0000000 --- a/source/elna/result.d +++ /dev/null @@ -1,107 +0,0 @@ -module elna.result; - -import std.typecons; -import tanya.container.array; -import tanya.container.string; - -/** - * Position in the source text. - */ -struct Position -{ - /// Line. - size_t line = 1; - - /// Column. - size_t column = 1; -} - -struct CompileError -{ - private string message_; - - private Position position_; - - @disable this(); - - /** - * Params: - * message = Error text. - * position = Error position in the source text. - */ - this(string message, Position position) @nogc nothrow pure @safe - { - this.message_ = message; - this.position_ = position; - } - - /// Error text. - @property string message() const @nogc nothrow pure @safe - { - return this.message_; - } - - /// Error line in the source text. - @property size_t line() const @nogc nothrow pure @safe - { - return this.position_.line; - } - - /// Error column in the source text. - @property size_t column() const @nogc nothrow pure @safe - { - return this.position_.column; - } -} - -struct Result(T) -{ - Nullable!CompileError error; - T result; - - this(T result) - { - this.result = result; - this.error = typeof(this.error).init; - } - - this(string message, Position position) - { - this.result = T.init; - this.error = CompileError(message, position); - } - - this(CompileError compileError) - { - this.result = null; - this.error = compileError; - } - - @disable this(); - - @property bool valid() const - { - return error.isNull; - } -} - -struct Reference -{ - enum Target - { - text, - high20, - lower12i - } - - String name; - size_t offset; - Target target; -} - -struct Symbol -{ - String name; - Array!ubyte text; - Array!Reference symbols; -} diff --git a/source/elna/riscv.d b/source/elna/riscv.d deleted file mode 100644 index 8fcfb2d..0000000 --- a/source/elna/riscv.d +++ /dev/null @@ -1,352 +0,0 @@ -module elna.riscv; - -import elna.extended; -import elna.ir; -import elna.result; -import std.algorithm; -import std.typecons; -import tanya.container.array; -import tanya.container.string; -import tanya.memory.allocator; - -enum XRegister : ubyte -{ - zero = 0, - ra = 1, - sp = 2, - gp = 3, - tp = 4, - t0 = 5, - t1 = 6, - t2 = 7, - s0 = 8, - s1 = 9, - a0 = 10, - a1 = 11, - a2 = 12, - a3 = 13, - a4 = 14, - a5 = 15, - a6 = 16, - a7 = 17, - s2 = 18, - s3 = 19, - s4 = 20, - s5 = 21, - s6 = 22, - s7 = 23, - s8 = 24, - s9 = 25, - s10 = 26, - s11 = 27, - t3 = 28, - t4 = 29, - t5 = 30, - t6 = 31, -} - -enum Funct3 : ubyte -{ - addi = 0b000, - slti = 0b001, - sltiu = 0b011, - andi = 0b111, - ori = 0b110, - xori = 0b100, - slli = 0b000, - srli = 0b101, - srai = 0b101, - add = 0b000, - slt = 0b010, - sltu = 0b011, - and = 0b111, - or = 0b110, - xor = 0b100, - sll = 0b001, - srl = 0b101, - sub = 0b000, - sra = 0b101, - beq = 0b000, - bne = 0b001, - blt = 0b100, - bltu = 0b110, - bge = 0b101, - bgeu = 0b111, - fence = 0b000, - fenceI = 0b001, - csrrw = 0b001, - csrrs = 0b010, - csrrc = 0b011, - csrrwi = 0b101, - csrrsi = 0b110, - csrrci = 0b111, - priv = 0b000, - sb = 0b000, - sh = 0b001, - sw = 0b010, - lb = 0b000, - lh = 0b001, - lw = 0b010, - lbu = 0b100, - lhu = 0b101, - jalr = 0b000, -} - -enum Funct12 : ubyte -{ - ecall = 0b000000000000, - ebreak = 0b000000000001, -} - -enum Funct7 : ubyte -{ - none = 0, - sub = 0b0100000 -} - -enum BaseOpcode : ubyte -{ - opImm = 0b0010011, - lui = 0b0110111, - auipc = 0b0010111, - op = 0b0110011, - jal = 0b1101111, - jalr = 0b1100111, - branch = 0b1100011, - load = 0b0000011, - store = 0b0100011, - miscMem = 0b0001111, - system = 0b1110011, -} - -struct Instruction -{ - private uint instruction; - - this(BaseOpcode opcode) @nogc - { - this.instruction = opcode; - } - @disable this(); - - ref Instruction i(XRegister rd, Funct3 funct3, XRegister rs1, uint immediate) - return scope @nogc - { - this.instruction |= (rd << 7) - | (funct3 << 12) - | (rs1 << 15) - | (immediate << 20); - - return this; - } - - ref Instruction s(uint imm1, Funct3 funct3, XRegister rs1, XRegister rs2, uint imm2 = 0) - return scope @nogc - { - this.instruction |= (imm1 << 7) - | (funct3 << 12) - | (rs1 << 15) - | (rs2 << 20) - | (imm2 << 25); - - return this; - } - - ref Instruction r(XRegister rd, Funct3 funct3, XRegister rs1, XRegister rs2, Funct7 funct7 = Funct7.none) - return scope @nogc - { - this.instruction |= (rd << 7) - | (funct3 << 12) - | (rs1 << 15) - | (rs2 << 20) - | (funct7 << 25); - - return this; - } - - ref Instruction u(XRegister rd, uint imm) - return scope @nogc - { - this.instruction |= (rd << 7) | (imm << 12); - - return this; - } - - ubyte[] encode() return scope @nogc - { - return (cast(ubyte*) (&this.instruction))[0 .. uint.sizeof]; - } -} - -class RiscVVisitor : IRVisitor -{ - Array!Instruction instructions; - bool registerInUse; - uint variableCounter = 1; - Array!Reference references; - - override void visit(Node) @nogc - { - } - - override void visit(Definition definition) @nogc - { - // Prologue. - this.instructions.insertBack( - Instruction(BaseOpcode.opImm) - .i(XRegister.sp, Funct3.addi, XRegister.sp, cast(uint) -32) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.store) - .s(28, Funct3.sw, XRegister.sp, XRegister.s0) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.store) - .s(24, Funct3.sw, XRegister.sp, XRegister.ra) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.opImm) - .i(XRegister.s0, Funct3.addi, XRegister.sp, 32) - ); - - foreach (statement; definition.statements[]) - { - statement.accept(this); - } - foreach (variableDeclaration; definition.variableDeclarations[]) - { - variableDeclaration.accept(this); - } - // Print the result. - this.instructions.insertBack( - Instruction(BaseOpcode.opImm) - .i(XRegister.a1, Funct3.addi, XRegister.a0, 0) - ); - this.references.insertBack(Reference(String(".CL0"), instructions.length * 4, Reference.Target.high20)); - this.instructions.insertBack( - Instruction(BaseOpcode.lui).u(XRegister.a5, 0) - ); - this.references.insertBack(Reference(String(".CL0"), instructions.length * 4, Reference.Target.lower12i)); - this.instructions.insertBack( - Instruction(BaseOpcode.opImm).i(XRegister.a0, Funct3.addi, XRegister.a5, 0) - ); - this.references.insertBack(Reference(String("printf"), instructions.length * 4, Reference.Target.text)); - this.instructions.insertBack( - Instruction(BaseOpcode.auipc).u(XRegister.ra, 0) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.jalr) - .i(XRegister.ra, Funct3.jalr, XRegister.ra, 0) - ); - // Set the return value (0). - this.instructions.insertBack( - Instruction(BaseOpcode.op) - .r(XRegister.a0, Funct3.and, XRegister.zero, XRegister.zero) - ); - - // Epilogue. - this.instructions.insertBack( - Instruction(BaseOpcode.load) - .i(XRegister.s0, Funct3.lw, XRegister.sp, 28) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.load) - .i(XRegister.ra, Funct3.lw, XRegister.sp, 24) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.opImm) - .i(XRegister.sp, Funct3.addi, XRegister.sp, 32) - ); - this.instructions.insertBack( - Instruction(BaseOpcode.jalr) - .i(XRegister.zero, Funct3.jalr, XRegister.ra, 0) - ); - } - - override void visit(Expression) @nogc - { - } - - override void visit(Statement statement) @nogc - { - statement.expression.accept(this); - } - - override void visit(Variable variable) @nogc - { - const freeRegister = this.registerInUse ? XRegister.a0 : XRegister.t0; - - // movl -x(%rbp), %eax; where x is a number. - this.instructions.insertBack( - Instruction(BaseOpcode.load) - .i(freeRegister, Funct3.lw, XRegister.sp, - cast(byte) (variable.counter * 4)) - ); - } - - override void visit(VariableDeclaration) @nogc - { - } - - override void visit(Number number) @nogc - { - const freeRegister = this.registerInUse ? XRegister.a0 : XRegister.t0; - - this.instructions.insertBack( - Instruction(BaseOpcode.opImm) // movl $x, %eax; where $x is a number. - .i(freeRegister, Funct3.addi, XRegister.zero, number.value) - ); - } - - override void visit(BinaryExpression expression) @nogc - { - this.registerInUse = true; - expression.lhs.accept(this); - this.registerInUse = false; - expression.rhs.accept(this); - - // Calculate the result and assign it to a variable on the stack. - final switch (expression.operator) - { - case BinaryOperator.sum: - this.instructions.insertBack( - Instruction(BaseOpcode.op) - .r(XRegister.a0, Funct3.add, XRegister.a0, XRegister.t0) - ); - break; - case BinaryOperator.subtraction: - this.instructions.insertBack( - Instruction(BaseOpcode.op) - .r(XRegister.a0, Funct3.sub, XRegister.a0, XRegister.t0, Funct7.sub) - ); - break; - } - this.instructions.insertBack( // movl %eax, -x(%rbp); where x is a number. - Instruction(BaseOpcode.store) - .s(cast(uint) (this.variableCounter * 4), Funct3.sw, XRegister.sp, XRegister.a0) - ); - - ++this.variableCounter; - } -} - -Symbol writeNext(Definition ast) @nogc -{ - Array!Instruction instructions; - Array!Reference references; - auto visitor = defaultAllocator.make!RiscVVisitor; - scope (exit) - { - defaultAllocator.dispose(visitor); - } - visitor.visit(ast); - - auto program = Symbol(String("main")); - - program.symbols = move(visitor.references); - foreach (ref instruction; visitor.instructions) - { - program.text.insertBack(instruction.encode); - } - return program; -} diff --git a/source/main.d b/source/main.d deleted file mode 100644 index dcb9fa0..0000000 --- a/source/main.d +++ /dev/null @@ -1,33 +0,0 @@ -import elna.backend; -import elna.ir; -import elna.arguments; -import std.path; -import std.sumtype; -import tanya.container.string; -import tanya.memory.allocator; -import tanya.memory.mmappool; - -int main(string[] args) -{ - defaultAllocator = MmapPool.instance; - - return Arguments.parse(args).match!( - (ArgumentError argumentError) => 4, - (Arguments arguments) { - String outputFilename; - if (arguments.output is null) - { - outputFilename = arguments - .inFile - .baseName - .withExtension("o"); - } - else - { - outputFilename = String(arguments.output); - } - - return generate(arguments.inFile, outputFilename); - } - ); -} diff --git a/tests/const_list.eln b/tests/const_list.eln deleted file mode 100644 index 375627f..0000000 --- a/tests/const_list.eln +++ /dev/null @@ -1,3 +0,0 @@ -const a = 1, b = 2; -! a + b -. diff --git a/tests/expectations/const_list.txt b/tests/expectations/const_list.txt deleted file mode 100644 index 00750ed..0000000 --- a/tests/expectations/const_list.txt +++ /dev/null @@ -1 +0,0 @@ -3 diff --git a/tests/expectations/left_nested_sum.txt b/tests/expectations/left_nested_sum.txt deleted file mode 100644 index 45a4fb7..0000000 --- a/tests/expectations/left_nested_sum.txt +++ /dev/null @@ -1 +0,0 @@ -8 diff --git a/tests/expectations/subtraction.txt b/tests/expectations/subtraction.txt deleted file mode 100644 index d00491f..0000000 --- a/tests/expectations/subtraction.txt +++ /dev/null @@ -1 +0,0 @@ -1 diff --git a/tests/expectations/sum.txt b/tests/expectations/sum.txt deleted file mode 100644 index 45a4fb7..0000000 --- a/tests/expectations/sum.txt +++ /dev/null @@ -1 +0,0 @@ -8 diff --git a/tests/expectations/sums.txt b/tests/expectations/sums.txt deleted file mode 100644 index 45a4fb7..0000000 --- a/tests/expectations/sums.txt +++ /dev/null @@ -1 +0,0 @@ -8 diff --git a/tests/left_nested_sum.eln b/tests/left_nested_sum.eln deleted file mode 100644 index aae2a77..0000000 --- a/tests/left_nested_sum.eln +++ /dev/null @@ -1,2 +0,0 @@ -! (3 + 4) + 1 -. diff --git a/tests/subtraction.eln b/tests/subtraction.eln deleted file mode 100644 index 00917c6..0000000 --- a/tests/subtraction.eln +++ /dev/null @@ -1,2 +0,0 @@ -! 5 - 4 -. diff --git a/tests/sum.eln b/tests/sum.eln deleted file mode 100644 index 6b4e720..0000000 --- a/tests/sum.eln +++ /dev/null @@ -1,2 +0,0 @@ -! 1 + 7 -. diff --git a/tests/sums.eln b/tests/sums.eln deleted file mode 100644 index 59586e9..0000000 --- a/tests/sums.eln +++ /dev/null @@ -1,2 +0,0 @@ -! 1 + (3 + 4) -.