1593

Committed 01 May 2026 12:00PM UTC coverage: 97.725% (-0.02%) from 97.744%

Build # 1593

Build Type

push

cirrus-ci

Committed by

ped7g

Commit Message

deb: fixes to make build package pass first time

Lua version 5.4 enforced as that's all I have in distro repo right now.
Should be probably depending on target distribution and their
availability.

Some file pruning and attempt to do the "right thing" like re-generating
docs/documentation.html during package build.

Resulting .deb packages were not vetted, just committing state when it
did pass, before checking content.

Coverage Stats

10136 of 10372 relevant lines covered (97.72%)

172270.29 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

99.41

/sjasm/reader.cpp

/*

  SjASMPlus Z80 Cross Compiler

  This is modified source of SjASM by Aprisobal - aprisobal@tut.by

  Copyright (c) 2006 Sjoerd Mastijn

  This software is provided 'as-is', without any express or implied warranty.
  In no event will the authors be held liable for any damages arising from the
  use of this software.

  Permission is granted to anyone to use this software for any purpose,
  including commercial applications, and to alter it and redistribute it freely,
  subject to the following restrictions:

  1. The origin of this software must not be misrepresented; you must not claim
         that you wrote the original software. If you use this software in a product,
         an acknowledgment in the product documentation would be appreciated but is
         not required.

  2. Altered source versions must be plainly marked as such, and must not be
         misrepresented as being the original software.

  3. This notice may not be removed or altered from any source distribution.

*/

// reader.cpp

#include "sjdefs.h"

//enum EDelimiterType          { DT_NONE, DT_QUOTES, DT_APOSTROPHE, DT_ANGLE, DT_COUNT };
static const char delimiters_b[] = { ' ',    '"',       '\'',          '<',      0 };
static const char delimiters_e[] = { ' ',    '"',       '\'',          '>',      0 };
static const std::array<EDelimiterType, 3> delimiters_all = {DT_QUOTES, DT_APOSTROPHE, DT_ANGLE};
static const std::array<EDelimiterType, 3> delimiters_noAngle = {DT_QUOTES, DT_APOSTROPHE, DT_COUNT};

int cmphstr(char*& p1, const char* p2, bool allowParenthesisEnd) {
        unsigned int i = 0;
        // check initial non-alpha chars without deciding the upper/lower case of test
        while (p2[i] && !isalpha((byte)p2[i])) {
                if (p1[i] != p2[i]) return 0;
                ++i;
        }
        // now the first alpha char will make whole test upper or lower case like.
        if (isupper((byte)p1[i])) {
                while (p2[i]) {
                        if (p1[i] != toupper((byte)p2[i])) return 0;
                        ++i;
                }
        } else {
                while (p2[i]) {
                        if (p1[i] != p2[i]) return 0;
                        ++i;
                }
        }
        if (p1[i]) {                // there is some character after the first word
                // whitespace, EOL-comment and block-comment-start keep the match valid
                // also starting parenthesis when allowParenthesisEnd
                if (
                        !White(p1[i]) && \
                        !(';' == p1[i]) && \
                        !('/' == p1[i] && '/' == p1[i+1]) && \
                        !('/' == p1[i] && '*' == p1[i+1]) && \
                        !(allowParenthesisEnd && '(' == p1[i])
                ) {
                        return 0;        // anything else invalidates the found match
                }
        }
        // space, tab, enter, \0, ... => "match"
        p1 += i;
        return 1;
}

bool White(const char c) {
        return c && (c&255) <= ' ';
}

bool White() {
        return White(*lp);
}

int SkipBlanks(char*& p) {
        while (White(*p)) ++p;
        return (*p == 0);
}

int SkipBlanks() {
        return SkipBlanks(lp);
}

void SkipParam(char*& p) {
        while (*p && (*p != ',')) ++p;
}

void SkipToEol(char*& p) {
        while (*p) ++p;
}

int NeedEQU() {
        char* olp = lp;
        SkipBlanks();
        /*if (*lp=='=') { ++lp; return 1; }*/
        /* cut: if (*lp=='=') { ++lp; return 1; } */
        if (*lp == '.') {
                ++lp;
        }
        if (cmphstr(lp, "equ")) {
                return 1;
        }
        lp = olp;
        return 0;
}

int NeedDEFL() {
        char* olp = lp;
        SkipBlanks();
        if (*lp == '=') {
                ++lp;
                return 1;
        }
        if (*lp == '.') {
                ++lp;
        }
        if (cmphstr(lp, "defl")) {
                return 1;
        }
        lp = olp;
        return 0;
}

bool NeedIoC() {
        SkipBlanks();
        if ('(' != lp[0] || 'c' != tolower((byte)lp[1]) || ')' != lp[2]) return false;
        lp += 3;
        return true;
}

bool isMacroNext() {        // checks if ".macro" directive is ahead (but doesn't consume it)
        if (SkipBlanks()) return false;
        char* p = lp;
        if ('.' == *p) ++p;
        return cmphstr(p, "macro");
}

bool anyComma(char*& p) {
        SkipBlanks(p);
        if (*p != ',') return false;
        ++p;
        return true;
}

bool comma(char*& p) {
        SkipBlanks(p);
        if (',' != p[0] || ',' == p[1]) return false;        // detect double-comma as FALSE state
        ++p;
        return true;
}

bool doubleComma(char* & p) {
        SkipBlanks(p);
        if (',' != p[0] || ',' != p[1]) return false;
        p += 2;
        return true;
}

bool nonMaComma(char* & p) {
        if (Options::syx.isMultiArgPlainComma()) return false;        // comma is also multi-arg => FALSE here
        return comma(p);
}

bool relaxedMaComma(char* & p) {
        SkipBlanks(p);
        if (',' != p[0]) return false;                // no comma
        if (',' == p[1]) {
                // double comma detected, accept it only in --syntax=a mode
                if (Options::syx.isMultiArgPlainComma()) return false;
                p += 2;
                return true;
        }
        // single comma is enough in relaxed way, even in --syntax=a mode
        ++p;
        return true;
}

//enum EBracketType          { BT_NONE, BT_ROUND, BT_CURLY, BT_SQUARE, BT_COUNT };
static const char brackets_b[] = { 0,      '(',      '{',      '[',       0 };
static const char brackets_e[] = { 0,      ')',      '}',      ']',       0 };
static int expectedAddressClosingBracket = -1;

// memory-address bracket opener (only "(" and "[" types supported)
EBracketType OpenBracket(char*& p) {
        SkipBlanks(p);
        if (2 == Options::syx.MemoryBrackets && brackets_b[BT_ROUND] == *p) return BT_NONE;                // disabled "()"
        for (const EBracketType bt : {BT_ROUND, BT_SQUARE}) {
                if (brackets_b[bt] == *p) {
                        expectedAddressClosingBracket = brackets_e[bt];
                        ++p;
                        return bt;
                }
        }
        expectedAddressClosingBracket = -1;
        return BT_NONE;
}

int CloseBracket(char*& p) {
        SkipBlanks(p);
        if (*p != expectedAddressClosingBracket) return 0;
        expectedAddressClosingBracket = -1;
        ++p;
        return 1;
}

char* ParenthesesEnd(char* p) {
        int depth = 0;        char pc;
        if (SkipBlanks(p) || '(' != *p) return nullptr;
        while (0 != (pc = *p++)) {
                if ('(' == pc) ++depth;
                else if (')' == pc && 0 == --depth) {
                        SkipBlanks(p);
                        return p;
                }
        }
        return nullptr;
}

char nidtemp[LINEMAX], *nidsubp = nidtemp;

//TODO v2.x: review GetID usage and make it more clear where are which characters legal
// add GetLabel where appropriate, handle "@" + "." modifiers more consistently and transparently)
char* GetID(char*& p) {
        char* np = nidtemp;
        if (SkipBlanks(p) || (!isLabelStart(p, false) && *p != '.')) return nullptr;
        while (islabchar((byte)*p)) *np++ = *p++;
        *np = 0;
        return nidtemp;
}

void ResetGrowSubId() {
        nidsubp = nidtemp;                        // reset temporary ID, starting new one
        *nidsubp = 0;
}

char* GrowSubId(char* & p) {        // appends next part of ID
        // The caller function ReplaceDefineInternal already assures the first char of ID is (isalpha() || '_')
        // so there are no extra tests here to verify validity of first character (like GetID(..) must do)
        const bool isSubwordSubstitution = Options::syx.IsSubwordSubstitution;        // help compiler -O2
        bool startsAtUnderscore = ('_' == *p);
        // add sub-parts delimiter in separate step (i.e. new ID grows like: "a", "a_", "a_b", ...)
        while (islabchar(*p)) {
                *nidsubp++ = *p++;
                // break at sub-word boundaries when new underscore block starts or ends
                if (isSubwordSubstitution && (('_' == *p) != startsAtUnderscore)) break;
        }
        if (nidtemp+LINEMAX <= nidsubp) Error("ID too long, buffer overflow detected.", NULL, FATAL);
        *nidsubp = 0;
        return nidtemp[0] ? nidtemp : nullptr;        // return non-empty string or nullptr
}

char* GrowSubIdByExtraChar(char* & p) {        // append the next char even if not a legal label/ID char
        // the caller function is responsible for all the validation, this just adds single char
        *nidsubp++ = *p++;
        if (nidtemp+LINEMAX <= nidsubp) Error("ID too long, buffer overflow detected.", NULL, FATAL);
        *nidsubp = 0;
        if (!nidtemp[0]) return NULL;        // result is empty string, return NULL rather
        return nidtemp;
}

char instrtemp[LINEMAX];

char* getinstr(char*& p) {
        char* np = instrtemp;
        SkipBlanks(p);
        if (!isalpha((byte)*p) && *p != '.') {
                return 0;
        } else {
                *np = *p; ++p; ++np;
        }
        while (*p) {
                if (!isalnum((byte)*p) && *p != '_') {
                        break;
                } /////////////////////////////////////
                *np = *p; ++p; ++np;
        }
        *np = 0;
        // colon after instruction can't happen, only if it was meant as label at beginning of line
        if (':' == *p) return nullptr;
        if (!Options::syx.CaseInsensitiveInstructions) return instrtemp;
        // lowercase the retrieved "instruction" string when option "--syntax=i" is used
        while (instrtemp <= --np) {
                *np = tolower((byte)*np);
        }
        return instrtemp;
}

int check8(aint val) {
        if (val < -256 || val > 255) {
                char buffer[64];
                SPRINTF2(buffer, 64, "value 0x%X is truncated to 8bit value: 0x%02X", val, val&0xFF);
                Warning(buffer);
                return 0;
        }
        return 1;
}

int check8o(aint val)
{
        if (val < -128 || val > 127) {
                char buffer[32];
                SPRINTF1(buffer, 32, "Offset out of range (%+i)", val);
                Error(buffer, nullptr, IF_FIRST);
                return 0;
        }
        return 1;
}

int check16(aint val) {
        if (val < -65536 || val > 65535) {
                char buffer[64];
                SPRINTF2(buffer, 64, "value 0x%X is truncated to 16bit value: 0x%04X", val, val&0xFFFF);
                Warning(buffer);
                return 0;
        }
        return 1;
}

int check16u(aint val) {
        if (val < 0 || val > 65535) {
                char buffer[64];
                SPRINTF2(buffer, 64, "value 0x%X is truncated to 16bit value: 0x%04X", val, val&0xFFFF);
                Warning(buffer);
                return 0;
        }
        return 1;
}

int check24(aint val) {
        if (val < -16777216 || val > 16777215) {
                char buffer[64];
                SPRINTF2(buffer, 64, "value 0x%X is truncated to 24bit value: 0x%06X", val, val&0xFFFFFF);
                Warning(buffer);
                return 0;
        }
        return 1;
}

void checkLowMemory(byte hiByte, byte lowByte) {
        if (hiByte || Relocation::type) return;
        // for addresses 0..255 issue warning
        WarningById(W_READ_LOW_MEM, lowByte);
}

int need(char*& p, char c) {
        SkipBlanks(p);
        if (*p != c) {
                return 0;
        }
        ++p; return 1;
}

int needa(char*& p, const char* c1, int r1, const char* c2, int r2, const char* c3, int r3, bool allowParenthesisEnd) {
        //  SkipBlanks(p);
        if (!isalpha((byte)*p)) {
                return 0;
        }
        if (cmphstr(p, c1, allowParenthesisEnd)) {
                return r1;
        }
        if (c2 && cmphstr(p, c2, allowParenthesisEnd)) {
                return r2;
        }
        if (c3 && cmphstr(p, c3, allowParenthesisEnd)) {
                return r3;
        }
        return 0;
}

int need(char*& p, const char* c) {
        SkipBlanks(p);
        while (*c) {
                if (*p != *c) {
                        c += 2; continue;
                }
                ++c;
                if (*c == ' ') {
                        ++p; return *(c - 1);
                }
                if (*c == '_' && *(p + 1) != *(c - 1)) {
                        ++p; return *(c - 1);
                }
                if (*(p + 1) == *c) {
                        p += 2; return *(c - 1) + *c;
                }
                ++c;
        }
        return 0;
}

int getval(int p) {
        assert(('0' <= p && p <= '9') || ('A' <= p && p <= 'Z') || ('a' <= p && p <= 'z'));
        if (p <= '9') return p - '0';
        return (p|0x20) - 'a' + 10;
}

const char* getNumericValueLastErr = NULL;
const char* const getNumericValueErr_syntax = "Syntax error";
const char* const getNumericValueErr_digit = "Digit not in base";
const char* const getNumericValueErr_no_digit = "Missing next digit";
const char* const getNumericValueErr_overflow = "Overflow";

bool GetNumericValue_ProcessLastError(const char* const srcLine) {
        if (NULL == getNumericValueLastErr) return false;
        Error(getNumericValueLastErr, srcLine, SUPPRESS);
        // Overflow type error lets assembler to emit truncated machine code (return "false" here)
        return (getNumericValueErr_overflow != getNumericValueLastErr);
}

bool GetNumericValue_TwoBased(char*& p, const char* const pend, aint& val, const int shiftBase) {
        if (shiftBase < 1 || 5 < shiftBase) Error("Internal error, wrong base", NULL, FATAL);
        getNumericValueLastErr = NULL;
        val = 0;
        if (pend <= p) {                // no actual digits between format specifiers
                getNumericValueLastErr = getNumericValueErr_syntax;
                return false;
        }
        aint digit;
        const int base = 1<<shiftBase;
        const aint overflowMask = (~0UL)<<(32-shiftBase);
        while (p < pend) {
                const byte charDigit = *p++;
                if (('\'' == charDigit || '_' == charDigit) && isalnum((byte)*p)) continue;
                if (0 == charDigit || !isalnum(charDigit)) {
                        getNumericValueLastErr = getNumericValueErr_no_digit;
                        break;
                }
                if (base <= (digit = getval(charDigit))) {
                        getNumericValueLastErr = getNumericValueErr_digit;
                        break;
                }
                if (val & overflowMask) getNumericValueLastErr = getNumericValueErr_overflow;
                val = (val<<shiftBase) + digit;
        }
        return (NULL == getNumericValueLastErr);
}

bool GetNumericValue_IntBased(char*& p, const char* const pend, aint& val, const int base) {
        if (base < 2 || 36 < base) Error("Internal error, wrong base", NULL, FATAL);
        getNumericValueLastErr = NULL;
        val = 0;
        if (pend <= p) {                // no actual digits between format specifiers
                getNumericValueLastErr = getNumericValueErr_syntax;
                return false;
        }
        aint digit;
        while (p < pend) {
                const byte charDigit = *p++;
                if (('\'' == charDigit || '_' == charDigit) && isalnum((byte)*p)) continue;
                if (0 == charDigit || !isalnum(charDigit)) {
                        getNumericValueLastErr = getNumericValueErr_no_digit;
                        break;
                }
                if (base <= (digit = getval(charDigit))) {
                        getNumericValueLastErr = getNumericValueErr_digit;
                        break;
                }
                const uint32_t oval = static_cast<uint32_t>(val);
                val = (val * base) + digit;
                if (static_cast<uint32_t>(val) < oval) getNumericValueLastErr = getNumericValueErr_overflow;
        }
        return (NULL == getNumericValueLastErr);
}

// parses number literals, forces result to be confined into 32b (even on 64b platforms,
// to have stable results in listings/tests across platforms).
int GetConstant(char*& op, aint& val) {
        // the input string has been already detected as numeric literal by ParseExpPrim
        assert(isdigit((byte)*op) || '#' == *op || '$' == *op || '%' == *op);
        // find end of the numeric literal (pointer is beyond last alfa/digit character
        char* pend = op;
        if ('#' == *pend || '$' == *pend || '%' == *pend) ++pend;
        while (isalnum((byte)*pend) || (('\'' == *pend || '_' == *pend) && isalnum((byte)pend[1]))) ++pend;
        char* const hardEnd = pend;
        bool has_decimal_part = ('.' == *hardEnd) && isalnum((byte)hardEnd[1]);
        // check if the format is defined by prefix (#, $, %, 0x, 0X, 0b, 0B, 0q, 0Q)
        char* p = op;
        int shiftBase = 0, base = 0;
        if ('#' == *p || '$' == *p) {
                shiftBase = 4;
                ++p;
        } else if ('0' == p[0] && 'x' == (p[1]|0x20)) {
                shiftBase = 4;
                p += 2;
        } else if ('0' == p[0] && 'b' == (p[1]|0x20) && 'h' != (pend[-1]|0x20) ) {
                shiftBase = 1;                // string 0b800h is hexadecimal, not binary (legacy compatibility)
                p += 2;
        } else if ('0' == p[0] && 'q' == (p[1]|0x20)) {
                shiftBase = 3;
                p += 2;
        } else if ('%' == *p) {
                shiftBase = 1;
                ++p;
        }
        // if the base is still undecided, check for suffix format specifier
        if (0 == shiftBase) {
                switch (pend[-1]|0x20) {
                        case 'h': --pend; shiftBase = 4;  has_decimal_part = false; break;
                        case 'q': --pend; shiftBase = 3;  has_decimal_part = false; break;
                        case 'o': --pend; shiftBase = 3;  has_decimal_part = false; break;
                        case 'b': --pend; shiftBase = 1;  has_decimal_part = false; break;
                        case 'd': --pend;      base = 10; has_decimal_part = false; break;
                        default:
                                base = 10;
                                break;
                }
        }
        if ('\'' == *p || '\'' == pend[-1] || '_' == *p || '_' == pend[-1]) {        // digit-group tick can't be first/last digit
                Error(getNumericValueErr_no_digit, op, SUPPRESS);
                return 0;
        }
        // parse the number into value
        if (0 < shiftBase) {
                if (!GetNumericValue_TwoBased(p, pend, val, shiftBase) && GetNumericValue_ProcessLastError(op))
                        return 0;
        } else {
                if (!GetNumericValue_IntBased(p, pend, val, base) && GetNumericValue_ProcessLastError(op))
                        return 0;
        }
        // check for possible decimal part and warn about it appropriately (can be user error or Lua string formatting)
        if (!has_decimal_part) {
                // no decimal part detected, all is done here
                op = hardEnd;
                return 1;
        }
        // possible decimal part detected, try to parse it just to throw it away (with warnings if enabled)
        p = hardEnd + 1;
        assert(isalnum((byte)*p));
        pend = hardEnd + 2;
        while (isalnum((byte)*pend) || (('\'' == *pend || '_' == *pend) && isalnum((byte)pend[1]))) ++pend;
        aint fractionVal;
        if (0 < shiftBase) {
                GetNumericValue_TwoBased(p, pend, fractionVal, shiftBase);
        } else {
                GetNumericValue_IntBased(p, pend, fractionVal, base);
        }
        // ignore overflow errors in fractional part, the value is thrown away any way, just report it as non-zero
        if (getNumericValueErr_overflow == getNumericValueLastErr) {
                fractionVal = 1;
                getNumericValueLastErr = nullptr;
        } else if (nullptr != getNumericValueLastErr) {                // but report other syntax errors
                GetNumericValue_ProcessLastError(hardEnd);
                return 0;
        }
        // warn about zero/non-zero fractional part in the numeral string
        WarningById(fractionVal ? W_NON_ZERO_DECIMAL : W_ZERO_DECIMAL, op);
        op = pend;
        return 1;
}

// parse single character of double-quoted string (backslash does escape characters)
int GetCharConstInDoubleQuotes(char*& op, aint& val) {
        if ('"' == *op || !*op) return 0;                // closing quotes or no more characters, return 0
        if ((val = *op++) != '\\') return 1;        // un-escaped character, just return it
        switch (val = *op++) {
        case '\\':
        case '\'':
        case '\"':
        case '\?':
                return 1;
        case '0':
                val = 0;
                return 1;
        case 'n':
        case 'N':
                val = 10;
                return 1;
        case 't':
        case 'T':
                val = 9;
                return 1;
        case 'v':
        case 'V':
                val = 11;
                return 1;
        case 'b':
        case 'B':
                val = 8;
                return 1;
        case 'r':
        case 'R':
                val = 13;
                return 1;
        case 'f':
        case 'F':
                val = 12;
                return 1;
        case 'a':
        case 'A':
                val = 7;
                return 1;
        case 'e':
        case 'E':
                val = 27;
                return 1;
        case 'd':
        case 'D':
                val = 127;
                return 1;
        default:
                break;
        }
        // return backslash as char value in case of unknown escape sequence
        // (to mimick older versions of sjasmplus like 1.07-1.10 behaviour)
        --op;
        val = '\\';
        Warning("Unknown escape", op-1);
        return 1;
}

// parse single character of apostrophe-quoted string (no escaping, double '' is apostrophe itself)
int GetCharConstInApostrophes(char*& op, aint& val) {
        if ('\'' == op[0] && '\'' == op[1]) {        // '' converts to actual apostrophe as value
                val = '\'';
                op += 2;
                return 1;
        }
        if ('\'' == *op || !*op) return 0;                // closing apostrophe or no more characters, return 0
        // normal character, just return it
        val = *op++;
        return 1;
}

// parse single/double quoted string literal as single value ('012' == 0x00303132)
int GetCharConst(char*& p, aint& val) {
        const char * const op = p;                // for error reporting
        char buffer[128];
        int bytes = 0, strRes;
        if (!(strRes = GetCharConstAsString(p, buffer, bytes))) return 0;                // no string detected
        val = 0;
        if (strRes < 0) return 0;                // some syntax/max_size error happened
        for (int ii = 0; ii < bytes; ++ii) val = (val << 8) + (255&buffer[ii]);
        if (0 == bytes) {
                Warning("Empty string literal converted to value 0!", op);
        } else if (4 < bytes) {
                const char oldCh = *p;
                *p = 0;                                                // shorten the string literal for warning display
                SPRINTF1(buffer, 128, "String literal truncated to 0x%X", val);
                Warning(buffer, op);
                *p = oldCh;                                        // restore it
        }
        return 1;
}

int GetBytes(char*& p, int e[], int add, int dc) {
        aint val;
        int t = 0, strRes;
        // reset alternate result flag in ParseExpression part of code
        Relocation::isResultAffected = false;
        do {
                const int oldT = t;
                char* const oldP = p;
                if (SkipBlanks(p)) {
                        Error("Expression expected", NULL, SUPPRESS);
                        break;
                }
                if (0 != (strRes = GetCharConstAsString(p, e, t, 128, add))) {
                        // string literal parsed (both types)
                        if (strRes < 0) break;                // syntax error happened
                        // single byte "strings" may have further part of expression, detect it here
                        if (1 == t - oldT && !SkipBlanks(p) && ',' != *p) {
                                // expression with single char detected (like 'a'|128), revert the string parsing
                                t = oldT;
                                p = oldP;                // and continue with the last code-path trying to parse expression
                        } else {        // string literal (not expression), handle the extra string literal logic
                                if (oldT < t && dc) {
                                        // mark last "string" byte with |128: single char in quotes *is* string
                                        // but single char in apostrophes *is not* (!) (no |128 then)
                                        int maxLengthNotString = (1 == strRes);                // 0 for quotes, 1 for apostrophes
                                        if (maxLengthNotString < (t - oldT)) e[t - 1] |= 128;
                                }
                                continue;
                        }
                }
                if (ParseExpressionNoSyntaxError(p, val)) {
                        check8(val);
                        Relocation::resolveRelocationAffected(t, Relocation::HIGH);
                        e[t++] = (val + add) & 255;
                } else {
                        Error("Syntax error", p, SUPPRESS);
                        break;
                }
        } while(comma(p) && t < 128);
        Relocation::checkAndWarn();
        e[t] = -1;
        if (t == 128 && *p) Error("Over 128 bytes defined in single DB/DC/... Values over", p, SUPPRESS);
        return t;
}

void GetStructText(char*& p, aint len, byte* data, const byte* initData) {
        assert(1 <= len && len <= CStructureEntry2::TEXT_MAX_SIZE && nullptr != data);
        // reset alternate result flag in ParseExpression part of code
        Relocation::isResultAffected = false;
        // "{}" is always required to keep the syntax less ambiguous
        // (prototype code was trying to be more relaxed, but it was quickly becoming too complicated)
        // (the relaxed sub-struct boundaries are confusing, eating "{}" a bit unexpectedly (to user))
        if (!need(p, '{')) {
                if (nullptr == initData) {
                        Error("TEXT field value must be enclosed in curly braces, missing '{'", p);
                } else {
                        memcpy(data, initData, len);
                }
                return;
        }
        aint ii = 0, val;
        do {
                // if no more chars/lines to be parsed, or ending curly brace incoming, finish the loop
                if (!PrepareNonBlankMultiLine(p) || '}' == *p) break;
                const int oldIi = ii;
                char* const oldP = p;
                int strRes;
                if (0 < (strRes = GetCharConstAsString(p, data, ii, len))) {
                        // string literal parsed (both types)
                        // single byte "strings" may have further part of expression, detect it here
                        if (1 == ii - oldIi && !SkipBlanks(p) && ',' != *p && '}' != *p) {
                                // expression with single char detected (like 'a'|128), revert the string parsing
                                ii = oldIi;
                                p = oldP;                // and continue with the last code-path trying to parse expression
                        } else {                        // string literal (not expression) parsed OK
                                continue;
                        }
                }
                if (-1 == strRes) break;                // syntax error happened
                if (-2 == strRes || len <= ii) {
                        Error("Maximum length of struct text reached. Values over", p, SUPPRESS);
                        break;
                }
                if (ParseExpressionNoSyntaxError(p, val)) {
                        check8(val);
                        data[ii++] = byte(val);
                } else {
                        Error("Syntax error", p, SUPPRESS);
                        break;
                }
        } while (comma(p));
        if (!PrepareNonBlankMultiLine(p) || !need(p, '}')) {
                Error("TEXT field value must be enclosed in curly braces, missing '}'", p);
                return;
        }
        Relocation::checkAndWarn();
        // some bytes were initialized explicitly
        if (nullptr != initData) {
                // init remaining bytes from initData
                while (ii < len) {
                        data[ii] = initData[ii];
                        ++ii;
                }
        } else {
                // init remaining bytes by last byte (or zero if none was defined)
                byte filler = 0 < ii ? data[ii - 1] : 0;
                while (ii < len) data[ii++] = filler;
        }
}

int GetBits(char*& p, int e[]) {
        EDelimiterType dt = DelimiterBegins(p, delimiters_noAngle);        //also skip blanks
        static int one = 0;                // the warning about multi-chars should be emitted only once per pass
        static bool zeroInDgWarning = false;
        int bytes = 0;
        while (*p && (dt == DT_NONE || delimiters_e[dt] != *p)) {
                if (128 <= bytes) {
                        Error("Over 128 bytes defined in DG. Bits over", p, SUPPRESS);
                        break;
                }
                // collect whole byte (eight bits)
                int value = 1, pch;
                while (value < 256 && *p && (pch = 255 & (*p++))) {
                        if (White(pch)) continue;                // skip spaces
                        value <<= 1;
                        if ('-' == pch || '.' == pch || '_' == pch) continue;
                        value |= 1;
                        if (LASTPASS != pass) continue;
                        if (0 < one && one != pch) {
                                Warning("[DG] multiple characters used for 'ones'");
                                one = -1;                                        // emit this warning only once
                        } else if (!one) one = pch;                // remember char used first time for "ones"
                        if ('0' == pch && !zeroInDgWarning) {
                                zeroInDgWarning = true;
                                Warning("[DG] character '0' in DG works as value 1");
                        }
                }
                if (value < 256) {                // there was not eight characters, ended prematurely
                        Error("[DG] byte needs eight characters", substitutedLine, SUPPRESS);
                } else {
                        e[bytes++] = value & 255;
                }
                SkipBlanks(p);
        }
        if (0 < one) one = 0;                // reset "ones" type if everything was OK this time
        e[bytes] = -1;
        if (dt == DT_NONE) return bytes;
        if (delimiters_e[dt] != *p)        Error("No closing delimiter", NULL, SUPPRESS);
        else                                                 ++p;
        return bytes;
}

int GetBytesHexaText(char*& p, int e[]) {
        const char* const op_full = p;
        int bytes = 0;
        do {
                EDelimiterType dt = DelimiterBegins(p, delimiters_noAngle);        //also skip blanks
                if (!*p) Error("no arguments");
                while (*p && (dt == DT_NONE || delimiters_e[dt] != *p)) {
                        const char* const op = p;
                        // collect whole byte = two hexa digits
                        aint val;
                        if (!GetNumericValue_TwoBased(p, p+2, val, 4) && GetNumericValue_ProcessLastError(op)) {
                                return 0;                // total failure, don't emit anything
                        }
                        if (128 <= bytes) {
                                Error("Over 128 bytes defined in DH/DEFH/HEX. Values over", op, SUPPRESS);
                                break;
                        }
                        e[bytes++] = val & 255;
                        SkipBlanks(p);                        // skip spaces
                        if (dt == DT_NONE && ',' == *p) break;        // loop through multi arguments in outer do-while loop
                }
                if (dt != DT_NONE) {
                        if (delimiters_e[dt] == *p)        {
                                ++p;
                                SkipBlanks(p);
                        } else Error("No closing delimiter", op_full, SUPPRESS);
                }
        } while (comma(p));
        e[bytes] = -1;
        return bytes;
}

// supports also string concatenation operator `..`, but then only the last delimiter is reported in result
delim_string_t GetDelimitedStringEx(char*& p) {
        delim_string_t result;
        do {
                result.second = DelimiterAnyBegins(p);
                const char deliE = delimiters_e[result.second];
                char *p_begin = p;
                while (*p && deliE != *p) ++p;
                result.first.append(p_begin, p);        // append found delimited string
                if (' ' != deliE) {
                        if (deliE == *p) {
                                ++p;
                        } else {
                                const char delimiterTxt[2] = { deliE, 0 };
                                Error("No closing delimiter", delimiterTxt, SUPPRESS);
                                result.first.clear();                // return "empty" string
                                return result;
                        }
                }
        } while (('.'+'.') == need(p, ".."));        // repeat when string concatenation operator is found
        return result;
}

std::string GetDelimitedString(char*& p) {
        return GetDelimitedStringEx(p).first;        // throw away delimiter type and return just string
}

bool isLabelStart(const char *p, bool modifiersAllowed) {
        if (modifiersAllowed) {
                if ('.' == *p || '@' == *p) return isLabelStart(p + 1, false);
        }
        return *p && (isalpha((byte)*p) || '_' == *p);
}

int islabchar(char p) {
        if (isalnum((byte)p) || p == '_' || p == '.' || p == '?' || p == '!' || p == '#' || p == '@') {
                return 1;
        }
        return 0;
}

EStructureMembers GetStructMemberId(char*& p) {
        if (*p == '#') {
                ++p;
                if (*p == '#') {
                        ++p;
                        return SMEMBALIGN;
                }
                return SMEMBBLOCK;
        }
        //  if (*p=='.') ++p;
        switch (*p * 2 + *(p + 1)) {
        case 'b'*2+'y':
        case 'B'*2+'Y':
                if (cmphstr(p, "byte")) return SMEMBBYTE;
                break;
        case 'w'*2+'o':
        case 'W'*2+'O':
                if (cmphstr(p, "word")) return SMEMBWORD;
                break;
        case 'b'*2+'l':
        case 'B'*2+'L':
                if (cmphstr(p, "block")) return SMEMBBLOCK;
                break;
        case 'd'*2+'b':
        case 'D'*2+'B':
                if (cmphstr(p, "db")) return SMEMBBYTE;
                break;
        case 'd'*2+'w':
        case 'D'*2+'W':
                if (cmphstr(p, "dw")) return SMEMBWORD;
                if (cmphstr(p, "dword")) return SMEMBDWORD;
                break;
        case 'd'*2+'s':
        case 'D'*2+'S':
                if (cmphstr(p, "ds")) return SMEMBBLOCK;
                break;
        case 'd'*2+'d':
        case 'D'*2+'D':
                if (cmphstr(p, "dd")) return SMEMBDWORD;
                break;
        case 'a'*2+'l':
        case 'A'*2+'L':
                if (cmphstr(p, "align")) return SMEMBALIGN;
                break;
        case 'd'*2+'e':
        case 'D'*2+'E':
                if (cmphstr(p, "defs")) return SMEMBBLOCK;
                if (cmphstr(p, "defb")) return SMEMBBYTE;
                if (cmphstr(p, "defw")) return SMEMBWORD;
                if (cmphstr(p, "defd")) return SMEMBDWORD;
                break;
        case 'd'*2+'2':
        case 'D'*2+'2':
                if (cmphstr(p, "d24")) return SMEMBD24;
                break;
        case 't'*2+'e':
        case 'T'*2+'E':
                if (cmphstr(p, "text")) return SMEMBTEXT;
                break;
        default:
                break;
        }
        return SMEMBUNKNOWN;
}

int GetMacroArgumentValue(char* & src, char* & dst) {
        SkipBlanks(src);
        const char* const dstOrig = dst, * const srcOrig = src;
        const char* dstStopTrim = dst;
        while (*src && ',' != *src) {
                // check if there is some kind of delimiter next (string literal or angle brackets expression)
                // the angle-bracket can only be used around whole argument (i.e. '<' must be first char)
                EDelimiterType delI = DelimiterBegins(src, srcOrig==src ? delimiters_all : delimiters_noAngle, false);
                // CPP numeric literals heuristic (eating digit-group apostrophes as regular digit char)
                if (DT_APOSTROPHE == delI && srcOrig < src) {
                        const char prevC = 0x20 | src[-1];        // also lowercase any ASCII letter
                        if (('0' <= prevC && prevC <= '9') || ('a' <= prevC && prevC <= 'f')) {
                                // ahead of apostrophe is character which can work as digit (up to hexa system)
                                // this heuristic doesn't bother to be super precise, because string-apostrophe
                                // should be on word boundary any way, not preceded by *any* digit of any base.
                                const char nextC = 0x20 | src[1];
                                if (('0' <= nextC && nextC <= '9') || ('a' <= nextC && nextC <= 'f')) {
                                        // by same logic, if also char after the apostrophe is any digit (up to base16)
                                        // turn the apostrophe from delimiter into digit-grouping char
                                        delI = DT_NONE;
                                }
                        }
                }
                if (DT_NONE == delI) {                // no delimiter found, ordinary expression, copy char by char
                        *dst++ = *src++;
                        continue;
                }
                // some delimiter found - parse those properly by their type
                if (DT_ANGLE != delI) *dst++ = *src;        // quotes are part of parsed value, angles are NOT
                ++src;                                                                        // advance over delimiter
                const char endCh = delimiters_e[delI];        // set expected ending delimiter
                while (*src) {
                        // handle escape sequences by the type of delimiter
                        switch (delI) {
                        case DT_ANGLE:
                                if (('!' == *src && '!' == src[1]) || ('!' == *src && '>' == src[1])) {
                                        *dst++ = src[1]; src += 2;        // escape sequence is converted into final char
                                        continue;
                                }
                                break;
                        case DT_QUOTES:
                                if ('\\' == *src && src[1]) {
                                        *dst++ = *src++;        *dst++ = *src++;
                                        continue;                                        // copy escape + escaped char (*any* non zero char)
                                }
                                break;
                        case DT_APOSTROPHE:
                                if ('\'' == *src && '\'' == src[1]) {
                                        *dst++ = *src++;        *dst++ = *src++;
                                        continue;                                        // copy two apostrophes (escaped apostrophe)
                                }
                                break;
                        default:        Error("Internal error. GetMacroArgumentValue()", NULL, FATAL);
                        }
                        if (endCh == *src) break;                        // ending delimiter found
                        *dst++ = *src++;                                        // just copy character
                }
                // ending delimiter must be identical to endCh
                if (endCh != *src) {
                        *dst = 0;                                                        // zero terminator of resulting string value
                        return 0;
                }
                // set ending delimiter for quotes and apostrophe (angles are stripped from value)
                if (DT_QUOTES == delI || DT_APOSTROPHE == delI) *dst++ = endCh;
                dstStopTrim = dst;                                                // should not trim right spaces beyond this point
                ++src;                                                                        // advance over delimiter
        }
        while (dstStopTrim < dst && White(dst[-1])) --dst;        // trim the right size space from value
        *dst = 0;                                                                        // zero terminator of resulting string value
        if (! *dstOrig) Warning("[Macro argument parser] empty value", srcOrig);
        return 1;
}

EDelimiterType DelimiterBegins(char*& src, const std::array<EDelimiterType, 3> delimiters, bool advanceSrc) {
        if ((0 == *src) || (advanceSrc && SkipBlanks(src))) return DT_NONE;
        for (const auto dt : delimiters) {
                if (delimiters_b[dt] != *src) continue;
                if (advanceSrc) ++src;
                return dt;
        }
        return DT_NONE;
}

EDelimiterType DelimiterAnyBegins(char*& src, bool advanceSrc) {
        return DelimiterBegins(src, delimiters_all, advanceSrc);
}

// returns (adjusts also "p" and "ei", and fills "e"):
//  -2 = buffer full
//  -1 = syntax error (missing quote/apostrophe)
//   0 = no string literal detected at p[0]
//   1 = (last) string literal in single quotes (apostrophe)
//   2 = (last) string literal in double quotes (")
template <class strT> int GetCharConstAsString(char* & p, strT e[], int & ei, int max_ei, int add) {
        if ('"' != *p && '\'' != *p) return 0;
        do {
                const int eiStart = ei;
                const char* const elementP = p;
                const bool quotes = ('"' == *p++);
                aint val;
                while (ei < max_ei && (quotes ? GetCharConstInDoubleQuotes(p, val) : GetCharConstInApostrophes(p, val))) {
                        e[ei++] = (val + add) & 255;
                }
                if ((quotes ? '"' : '\'') != *p) {        // too many/invalid arguments or zero-terminator can lead to this
                        if (*p) return -2;                                // too many arguments
                        Error("Syntax error", elementP, SUPPRESS);        // zero-terminator
                        return -1;
                }
                ++p;
                if ('Z' == *p) {
                        if (max_ei <= ei) return -2;        // no space for zero byte (keep p pointing at Z to report error)
                        ++p;
                        e[ei++] = 0;
                } else if ('C' == *p) {
                        ++p;
                        if (eiStart == ei) {
                                Error("C suffix can't patch empty string", elementP, SUPPRESS);
                                return -1;                        // empty string can't have last char patched
                        }
                        e[ei - 1] |= 0x80;
                }
                if (eiStart == ei) WarningById(W_EMPTY_STRING, elementP);
                if (('.'+'.') != need(p, "..")) return 1 + quotes;        // type of quotes reported for last segment only
                // string concatenation operator handling
                if (SkipBlanks(p) || ('"' != *p && '\'' != *p)) {
                        Error("[..] String literal expected", p, SUPPRESS);
                        return -1;        // another string was expected to follow
                }
        } while (true);
}

// make sure both specialized instances for `char` and `int` are available for whole app
// moved to end of file as it seems to derail test coverage data (either this, or the template code)
template int GetCharConstAsString<char>(char* & p, char e[], int & ei, int max_ei, int add);
template int GetCharConstAsString<int>(char* & p, int e[], int & ei, int max_ei, int add);

//eof reader.cpp

z00m128 / sjasmplus / 1593

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous