kerep/src/DtsodParser/DtsodV24_deserialize.c

#include "DtsodV24.h"
#include "../String/StringBuilder.h"

#define ARR_SZ_START 64


typedef struct DeserializeSharedData{
    const char* sh_text_first;
    char* sh_text;
    bool sh_partOfDollarList;
    bool sh_calledRecursively;
    allocator_ptr sh_tmp_al;
} DeserializeSharedData;

#define text shared->sh_text
#define partOfDollarList shared->sh_partOfDollarList
#define calledRecursively shared->sh_calledRecursively
#define tmp_al shared->sh_tmp_al


// special func for throwing error messages about wrong characters in deserializing text
Maybe ERROR_WRONGCHAR(const char c, char* _text, char* text_first, const char* srcfile, i32 line, const char* funcname){
    char errBuf[68];
    for(u8 n=0; n<sizeof(errBuf);n++)
        errBuf[n]='\0';
    char* errText=_text-32;
    u8 cplace=32;
    if(errText<text_first) {
        cplace=_text-text_first;
        errText=text_first;
    }
    u8 i=0;
    for(; i<cplace; i++){
        // writes 32 chars before the wrongchar
        errBuf[i]=errText[i];
    }
    //prints wrongchar in braces
    errBuf[i++]='<';
    errBuf[i++]=c;
    errBuf[i++]='>';
    for(; i<cplace+3+32; i++){
        // writes 32 chars after the wrongchar
        char _c=errText[i-2];
        errBuf[i]=_c;
        if(!_c) break;
    }
    char errmsg[1024];
    sprintf_s(errmsg, sizeof(errmsg), "unexpected <%c> at:\n"
                    "  \"%s\"\n"
                    "\\___[%s:%d] %s()",
                    c,errBuf, srcfile,line,funcname);
    safethrow(errmsg,;);
}
#define safethrow_wrongchar(C, freeMem) { freeMem; return ERROR_WRONGCHAR(C, text, shared->sh_text_first, __FILE__,__LINE__,__func__); }


Maybe __SkipComment(DeserializeSharedData* shared) {
    char c;

    while ((c=*++text) != '\n')
        if (!c) safethrow(ERR_ENDOFSTR,;);

    return MaybeNull;
}
#define SkipComment() __SkipComment(shared)

Maybe __ReadName(DeserializeSharedData* shared){
    char c;
    string nameStr={text+1,0};

    while ((c=*++text)) switch (c){
        case ' ':  case '\t':
        case '\r': case '\n':
            if(nameStr.length!=0)
                safethrow_wrongchar(c,;);
            nameStr.ptr++;
            break;
        case '=':  case ';':
        case '\'': case '"':
        case '[':  case ']':
        case '{':
            safethrow_wrongchar(c,;);
        case '#': ;
            char _c=c;
            char* _text=text;
            try(SkipComment(),_,;);
            if(nameStr.length!=0){
                text=_text;
                safethrow_wrongchar(_c,;);
            }
            nameStr.ptr=text+1; // skips '\n'
            break;
        case '}':
            if(!calledRecursively || nameStr.length!=0)
                safethrow_wrongchar(c,;);
            return SUCCESS(UniHeapPtr(char,NULL));
        case ':':
            return SUCCESS(UniHeapPtr(char,string_extract(tmp_al, nameStr)));
        case '$':
            if(nameStr.length!=0)
                safethrow_wrongchar(c,;);
            nameStr.ptr++;
            partOfDollarList=true;
            break;
        default:
            nameStr.length++;
            break;
    }

    if(nameStr.length>0) safethrow(ERR_ENDOFSTR,;);
    return SUCCESS(UniHeapPtr(char,NULL));
}
#define ReadName() __ReadName(shared)

Maybe __deserialize(char** _text, bool _calledRecursively, allocator_ptr _tmp_al);
Maybe __ReadValue(DeserializeSharedData* shared, bool* readingList);
#define ReadValue(rL) __ReadValue(shared, rL)

// returns part of <text> without quotes
Maybe __ReadString(DeserializeSharedData* shared){
    char c;
    bool prevIsBackslash=false;
    StringBuilder _sb;
    StringBuilder* b=&_sb;
    StringBuilder_construct(b, tmp_al);

    while ((c=*++text)){
        if(c=='"') {
            if(prevIsBackslash) {
                // replacing <\"> with <">
                StringBuilder_rmchar(b);
                StringBuilder_append_char(b,c);
                prevIsBackslash=false;
            }
            else {
                char* str=StringBuilder_build(b).ptr;
                return SUCCESS(UniHeapPtr(char,str));
            }
        }
        else {
            prevIsBackslash= c=='\\' && !prevIsBackslash;
            StringBuilder_append_char(b,c);
        }
    }

    safethrow(ERR_ENDOFSTR, StringBuilder_destruct(b));
}
#define ReadString() __ReadString(shared)

Maybe __ReadList(DeserializeSharedData* shared){
    Autoarr(Unitype) list;
    Autoarr_construct(&list, Unitype, ARR_SZ_START, NULL);
    bool readingList=true;
    while (true){
        try(ReadValue((&readingList)), m_val, Autoarr_destruct(&list))
            Autoarr_add(&list, m_val.value);
        if (!readingList){
            if(Unitype_isUniNull(m_val.value))
                Autoarr_pop(&list);
            break;
        }
    }

    return SUCCESS(UniHeapPtr(Autoarr_Unitype, &list));
};
#define ReadList() __ReadList(shared)

Maybe __ParseValue(DeserializeSharedData* shared, string str){
    const string trueStr={"true",4};
    const string falseStr={"false",5};

    switch(str.ptr[str.length-1]){
        // Bool
        case 'e':
            if(string_compare(str,trueStr))
                return SUCCESS(UniTrue);
            else if(string_compare(str,falseStr))
                return SUCCESS(UniFalse);
            else safethrow_wrongchar(*str.ptr,;);
        // Float64
        case 'f': {
            char* _c=string_extract(tmp_al, str);
            Unitype rez=UniFloat64(strtod(_c,NULL));
            // allocator_free(tmp_al,_c);
            return SUCCESS(rez);
        }
        // UInt64
        case 'u': {
            u64 lu=0;
            char* _c=string_extract(tmp_al, str);
            if(sscanf(_c, IFWIN("%llu", "%lu"), &lu)!=1){
                char err[64];
                sprintf_s(err, sizeof(err), "can't parse to int: <%s>", _c);
                safethrow(err, /*allocator_free(tmp_al, _c)*/);
            }
            // allocator_free(tmp_al, _c);
            return SUCCESS(UniUInt64(lu));
        }
        // Int64
        case '0': case '1': case '2': case '3': case '4':
        case '5': case '6': case '7': case '8': case '9': {
            i64 li=0;
            char* _c=string_extract(tmp_al, str);
            if(sscanf(_c, IFWIN("%lli", "%li"), &li)!=1){
                char err[64];
                sprintf_s(err, sizeof(err),"can't parse to int: <%s>",_c);
                // safethrow(err,allocator_free(tmp_al, _c));
            }
            // allocator_free(tmp_al, _c);
            return SUCCESS(UniInt64(li));
        }
        // wrong type
        default:
            safethrow_wrongchar(str.ptr[str.length-1],;);
    }

    safethrow(ERR_ENDOFSTR,;);
};
#define ParseValue(str) __ParseValue(shared, str)

Maybe __ReadValue(DeserializeSharedData* shared, bool* readingList){
    char c;
    string valueStr={text+1,0};
    Unitype value=UniNull;
    bool spaceAfterVal=false;
    
    while ((c=*++text)) switch (c){
        case ' ':  case '\t':
        case '\r': case '\n':
            if(valueStr.length!=0)
                spaceAfterVal=true;
            else valueStr.ptr++;
            break;
        case '=': case ':':
        case '}': case '$':
        case '\'':
            safethrow_wrongchar(c,Unitype_destruct(&value));
        case '#':;
            char _c=c;
            char* _text=text;
            try(SkipComment(),_,;);
            if(valueStr.length!=0){
                text=_text;
                safethrow_wrongchar(_c,Unitype_destruct(&value));
            }
            valueStr.ptr=text+1; // skips '\n'
            break;
        case '"':
            if(valueStr.length!=0) safethrow_wrongchar(c,Unitype_destruct(&value));
            try(ReadString(),maybeString,;)
                value=maybeString.value;
            break;
        case '{':
            if(valueStr.length!=0) safethrow_wrongchar(c,Unitype_destruct(&value));
            ++text; // skips '{'
            try(__deserialize(&text,true,tmp_al), val, Unitype_destruct(&value))
                value=val.value;
            break;
        case '[':
            if(valueStr.length!=0) safethrow_wrongchar(c,Unitype_destruct(&value));
            try(ReadList(),maybeList,Unitype_destruct(&value))
                value=maybeList.value;
            break;
        case ']':
            if(!readingList) safethrow_wrongchar(c,Unitype_destruct(&value));
            *readingList=false;
            goto return_value;
        case ';':
        case ',':
            return_value:
            if(valueStr.length!=0){
                if(!Unitype_isUniNull(value))
                    safethrow_wrongchar(c,Unitype_destruct(&value));
                try(ParseValue(valueStr),maybeParsed,;)
                    value=maybeParsed.value;
            }
            return SUCCESS(value);
        default:
            if(spaceAfterVal)
                safethrow_wrongchar(c,Unitype_destruct(&value));
            valueStr.length++;
            break;
    }

    safethrow(ERR_ENDOFSTR,;);
}


Maybe __deserialize(char** _text, bool _calledRecursively, allocator_ptr _tmp_al) {
    DeserializeSharedData _shared={
        .sh_text_first=*_text,
        .sh_text=*_text,
        .sh_partOfDollarList=false,
        .sh_calledRecursively=_calledRecursively,
        .sh_tmp_al=_tmp_al
    };
    DeserializeSharedData* shared=&_shared;
    Hashtable* dict=Hashtable_create();
    
    text--;
    while(true){
        try(ReadName(), maybeName, Hashtable_destruct(dict))
            if(!maybeName.value.VoidPtr) // end of file or '}' in recursive call
                goto END;
        char* nameCPtr=maybeName.value.VoidPtr;
        try(ReadValue(NULL), val, {
            Hashtable_destruct(dict);
            // do not use, free call order is incorrect
            // allocator_free(tmp_al, nameCPtr);
        }) {
            if(partOfDollarList){
                partOfDollarList=false;
                Autoarr(Unitype)* list;
                Unitype lu;
                if(Hashtable_tryGet(dict,nameCPtr, &lu)){
                    list=(Autoarr(Unitype)*)lu.VoidPtr;
                    // Key is not used in that case, because it is already added
                    // to the table with the first dollar list item.
                    // do not use, free call order is incorrect
                    // allocator_free(tmp_al, nameCPtr);
                }
                else{
                    list=allocator_alloc(tmp_al, sizeof(*list));
                    Autoarr_construct(list, Unitype, ARR_SZ_START, NULL);
                    Hashtable_add(dict,nameCPtr,UniHeapPtr(Autoarr_Unitype,list));
                }
                Autoarr_add(list,val.value);
            }
            else Hashtable_add(dict,nameCPtr,val.value);
        }
    }

    END:
    *_text=text;
    return SUCCESS(UniHeapPtr(Hashtable,dict));
}

Maybe DtsodV24_deserialize(char* _text) {
    LinearAllocator _tmp_al;
    LinearAllocator_construct(&_tmp_al, 1024);
    Maybe m=__deserialize(&_text, false, (allocator_ptr)&_tmp_al);
    LinearAllocator_destruct(&_tmp_al);
    return m;
}