在不同环境中运行C程序时出现分段错误

Question

I've recently started programming in C and I'm having trouble submitting a homework on a site called run.codes . . I suspected it had something to do with the C version that the site is using, so I tried to make my program compatible with older standards (C99), but it still not working.

Of all the online compilers I've tested, only this one gave me the same error as run.codes. In all the others, my program worked as expected. Since the "segmentation fault" error usually indicates that an unallowed attempt to access memory has ocurred, I've tried using Valgrind to track possible errors. My program happens to have several, even though it compiles and works fine in most environments.

Below is my program. I apologize the over-commenting and for the use of portuguese. My professor demands it to be that way.

#include <stdio.h>
#include <string.h>
#include <stdlib.h>


/*
 * Definicao de constantes. 
 */
//Tamanhos limites (por seguranca, sao usados valores maiores do que o necessario!)
#define MAX_CHARS 600
#define MAX_SENTENCES 100

//Tipos de perguntas
enum {
    POSITIVOS = 101, NEGATIVOS = 102, PALAVRA = 103
};

//Listas de palavras
const char *POSITIVE_WORDS[] = {"bom", "maravilhoso", "otimo", "sensacional", "excelente", "adorei", "gostei", "amei", "eficiente", "boa", "maravilhosa", "otima"};
const char *NEGATIVE_WORDS[] = {"detestei", "odiei", "ruim", "pessimo", "terrivel", "raiva", "odio", "pessima", "lento", "lenta", "fragil", "desisti"};
const char *INVERSION_WORDS[] = {"nao", "jamais", "nunca"};


/*
 * Implementacao de funcoes nao existentes no C99.
 */
//strdup (copia uma string)
char *my_strdup(char *src)
{
    size_t len = strlen(src) + 1; //o +1 serve para que o simbolo \0 seja copiado tambem
    char *s = malloc(len);

    if (s == NULL)
        return NULL;

    return (char *) memcpy(s, src, len); //memcpy retornara um ponteiro para s
}
//strsep (separa uma string usando um delimitador)
char *my_strsep(char** stringp, const char* delim)
{
  char *p, *start = *stringp;
  p = (start != NULL) ? strpbrk(start, delim) : NULL;

  if (p == NULL) {
    *stringp = NULL;
  }
  else {
    *p = '\0';
    *stringp = p + 1;
  }

  return start;
}


/*
 * Estrutura de um produto.
 */
typedef struct
{
    char *name; //nome do produto
    char sentences[MAX_SENTENCES][MAX_CHARS]; //sentencas que se referem ao produto
    int sentences_count; //numero de sentencas que se referem ao produto    
} Product;


/*
 * Associa uma sentenca a um produto.
 *
 * Product *p: produto em questao.
 * char *sentence: string contendo a sentenca.
 */
void addSentenceToProduct(Product *p, char *sentence)
{
    if(p->sentences_count < MAX_SENTENCES) {
        strcpy(p->sentences[p->sentences_count], sentence);
        p->sentences_count++;
    }
}


/*
 * Remove o ultimo caractere de uma string caso ele seja igual a c.
 *
 * char *str: a string que deseja-se alterar.
 * char c: o caractere que deve estar na ultima posicao da string.
 */
void removeLastChar(char *str, char c)
{
    int len = strlen(str);
    if(len > 0 && str[len-1] == c)
        str[len-1] = '\0';
}


/*
 * Le sentencas com multiplas palavras e as armazena na matriz dada.
 *
 * int count: quantidade de sentencas.
 * char sentencas[][]: matriz onde as sentencas serao armazenadas.
 */
void readSentences(int count, char sentences[][MAX_CHARS])
{
    for(int i = 0; i < count; i++) {
        fgets(sentences[i], MAX_CHARS, stdin);

        //ignora a entrada caso ela seja apenas um line breaker
        if(sentences[i][0] == '\n') {
            i--;
            continue;
        }

        //remove o line breaker do final da string, caso haja um
        removeLastChar(sentences[i], '\n');
    }
}


/*
 * Le sentencas na qual um produto é especificado no inicio. A funcao strsep() é usada para separar o produto do restante da sentenca.
 *
 * int count: quantidade de sentencas.
 * Product products[]: lista com os produtos
 * 
 */
int readSentencesWithProducts(int count, Product products[])
{
    char temp[MAX_SENTENCES][MAX_CHARS];
    readSentences(count, temp); //lendo a frase completa

    int productsCount = 0;
    for(int i = 0; i < count; i++) {
        char *temp2 = my_strdup(temp[i]);

        char *productName = my_strsep(&temp2, ";");
        removeLastChar(productName, ' '); //remove o espaco no final do nome do produto

        //Verifica se o produto ja esta na lista
        int index = -1; 
        for(int j = 0; j < productsCount; j++) {
            if(strcmp(productName, products[j].name) == 0) {
                index = j;
            }
        }

        //Novo produto, caso nao haja nenhum com o nome especificado
        if(index < 0) {
            index = productsCount;
            productsCount++;

            products[index].name = productName;
            products[index].sentences_count = 0;
        }

        //Descobre a sentenca e a adiciona ao produto associado
        char *sentence = my_strsep(&temp2, ";"); //separando sentencas
        addSentenceToProduct(&products[index], sentence);

        free(temp2); //liberando a memoria alocada com o strdup()
    }

    return productsCount;
}


/*
 * Retorna o index (referente a array de produtos) do produto ao qual a pergunta se refere. Retorna -1 caso o produto nao seja encontrado.
 *
 * char *question: referencia para a string que contem a pergunta.
 * Products products[]: array que contem os produtos.
 * int count: numero de produtos
 */
int findProductIndex(char *question, Product products[], int count)
{
    for(int i = 0; i < count; i++) {
        if(strstr(question, products[i].name) != NULL)
            return i;
    }

    return -1;
}


/*
 * Busca pela intencao da pergunta (o que ela quer saber).
 *
 * char *question: referencia para a string contendo a pergunta.
 */
int getQuestionIntent(char *question) 
{
    if(strstr(question, "palavra") != NULL)
        return PALAVRA;
    if(strstr(question, "negativos") != NULL)
        return NEGATIVOS;

    return POSITIVOS;
}


/*
 * Retorna a quantidade de vezes que uma dada palavra aparece em uma sentenca.
 *
 * char *word: palavra cujo uso sera contado.
 * char *sentence: sentenca que sera analisada.
 */
int countWordUsage(const char *word, char *sentence)
{
    int counter = 0;

    char *temp = my_strdup(sentence);
    char *resultWord = my_strsep(&temp, " ");

    while(resultWord != NULL) {
        if(strcmp(word, resultWord) == 0) 
            counter++;

        resultWord = my_strsep(&temp, " ");
    }       

    free(temp);
    return counter;
}


/*
 * Retorna 1 caso a palavra dada seja um inversor e 0 caso nao.
 *
 * char *word: palavra que sera checada.
 */
int isInversor(char *word)
{
    for(int i = 0; i < 3; i++) {
        if(strcmp(word, INVERSION_WORDS[i]) == 0)
            return 1;
    }
    return 0;
}


/*
 * Retorna 1 caso a palavra dada pertenca a classe das "positivas".
 *
 * char *word: palavra que sera analisada.
 */
int isPositive(char *word) {
    for(int i = 0; i < 12; i++) {
        if(strcmp(word, POSITIVE_WORDS[i]) == 0)
            return 1;
    }
    return 0;
}


/*
 * Retorna 1 caso a palavra dada pertenca a classe das "negativas".
 *
 * char *word: palavra que sera analisada.
 */
int isNegative(char *word) {
    for(int i = 0; i < 12; i++) {
        if(strcmp(word, NEGATIVE_WORDS[i]) == 0)
            return 1;
    }
    return 0;
}


/*
 * Retorna a quantidade de vezes que palavras de uma dada lista, levando-se em consideracao inversores, aparecem nas sentencas de um produto.
 *
 * Product p: produto que sera analisado.
 * int dividend: o dividendo da razao (POSITIVO ou NEGATIVO)
 */
double countPosNegRatio(Product p, const int dividend)
{
    int pos = 0, neg = 0, lastInversor = 4;
    for(int i = 0; i < p.sentences_count; i++) 
    {
        char *sentence = my_strdup(p.sentences[i]);
        char *word = my_strsep(&sentence, " ");

        int countedPos = 0, countedNeg = 0; //verifica se ja foi contada uma opiniao positiva ou negativa na sentenca em questao    
        while(word != NULL) 
        {
            if(isInversor(word)) {
                lastInversor = 1;
            }
            else {
                if(lastInversor > 3) {
                    //Positiva
                    if(isPositive(word) && !countedPos) {
                        pos++;
                        countedPos = 1;
                    }
                    //Negativa
                    else if(isNegative(word) && !countedNeg) {
                        neg++;
                        countedNeg = 1;
                    }
                }
                else {
                    //Negativa (inversao)
                    if(isPositive(word) && !countedNeg) {
                        neg++;
                        countedNeg = 1;
                    }
                    //Positiva(inversao)
                    else if(isNegative(word) && !countedPos) {
                        pos++;
                        countedPos = 1;
                    }
                }
                lastInversor++;
            }
            word = my_strsep(&sentence, " ");
        }
        free(sentence);
    }

    int total = pos + neg;
    return (dividend == POSITIVOS) ? (double) pos/total : (double) neg/total;
}


/*
 * Processa e responde a uma pergunta.
 *
 * char *question: pergunta que deve ser respondida.
 * Products products[]: lista de produtos
 * int count: quantidade total de produtos
 */
void processQuestion(char *question, Product products[], int count)
{
    int pIndex = findProductIndex(question, products, count);

    if(pIndex >= 0) 
    {
        Product p = products[pIndex];
        int intent = getQuestionIntent(question);

        switch(intent)
        {
            case POSITIVOS:
            case NEGATIVOS:
            {
                printf("%.1lf%%\n", 100*countPosNegRatio(p, intent));
                break;
            }
            case PALAVRA:
            {

                //descobre a palavra que se deseja buscar
                char *temp = my_strdup(question);
                my_strsep(&temp, " "); //descarta a primeira palavra
                char *word = my_strsep(&temp, " "); //armazena a segunda palavra

                //checa quantas vezes a palavra apareceu
                int counter = 0;
                for(int i = 0; i < p.sentences_count; i++) {
                    counter += countWordUsage(word, p.sentences[i]);
                }

                printf("%d\n", counter);
                break;
            }
            default:
                printf("Nao foi possivel detectar a intencao da pergunta!\n");
        }
    }
    else {
        printf("Produto nao encontrado!\n");
    }
}


/*
 * MAIN
 */
int main()
{
    int sentences_count, questions_count; 
    scanf("%d %d", &sentences_count, &questions_count);

    Product products[MAX_SENTENCES]; //armazena os produtos e suas respectivas sentencas
    int products_count = readSentencesWithProducts(sentences_count, products);

    char questions[MAX_SENTENCES][MAX_CHARS]; //armazena as perguntas
    readSentences(questions_count, questions);

    for(int i = 0; i < questions_count; i++) {
        processQuestion(questions[i], products, products_count);
    }


    /*
    //DEBUG
    for(int i = 0; i < productsCount; i++) {
        Product p = products[i];
        printf("-> %s:\n", p.name);
        for(int j = 0; j < p.sentences_count; j++) {
            printf("    %d - %s\n", j, p.sentences[j]); 
        }
    }
    */

    return 0;
}

The program is supposed to read comments about certain products and then answer some questions, like "how many times was the word X used?" or "what's the percentage of positive comments about product Y?" . The two first integers given as input specify, respectively, the number of comments and the number of questions. The rest of the input consists on all the comments followed by all the questions. Each comment is formatted this way: "name of the product ; comment" . The output consists of an answer for each of the questions.

Here is the log generated by Valgrind when I use the input below:

3 1
Televisor LG ; adorei o televisor lg ele e muito funcional Ø
Laptop HP ; nao gostei desse notebook pois ele esquenta muito Ø
Celular Motorola ; o formato do celular e muito bom bom mesmo bom demais Ø
palavra bom em Celular Motorola

The expected output (3), is given by the program. I'm using the following command to compile my code:

gcc trab.c -o trab -Wall -ggdb3

And the following parameters to initialize Valgrind:

valgrind --leak-check=full --show-leak-kinds=all --track-origins=yes --log-file=valgrind-out.txt ./trab

Finally, having given you all this information, my question is: I hate that the post ended up this big, but since the problem is weird, I've felt that I should give all the information I could. Thanks for your time!

my operational system is Ubuntu 19.04 and these ( here , here , here and here ) are some of the online compilers in which my program has run successfully. And this is the online compiler in which I get a segmentation fault error.

Answer 1

With the help of a friend, I've managed to solve my problem! I'm posting the solution here in case someone runs into a similar problem in the future.

. My program was simply using too much memory. To solve that, I've changed the size of the products array to a reasonable one (I was using much more than what I actually needed!). Look below:

#define MAX_SENTENCES 32 //It was 100 before!

Note that even though the number of errors detected by Valgrind has decreased considerably with this, there are still some memory leaks in my program!

This, unfortunately, wasn't enough for run.codes to accept my program. Some operational systems use \\r as the new line character (usually \\n ). I was using the function fgets() to read strings from the stdin. It also reads new line characters, in which I'm not interested, so I checked for them so they could be discarded. However, :

void readSentences(int count, char sentences[][MAX_CHARS]) {
        //...
        fgets(sentences[i], MAX_CHARS, stdin);
        if(sentences[i][0] == '\n') {
            i--;
            continue;
        }
        //...
}

When :

if(sentences[i][0] == '\n' || sentences[i][0] == '\r') //{ ... }

This was the cause of the "Segmentation Fault" error on some environments. After solving that, my program worked fine on them!