Compare commits

base: retoor:94e5d7bfa5d8444a4d3eb9772bf52ce0e60344ed
Branches Tags
retoor:main
..
compare: retoor:8e4a8f6f09fd766a320aaa49659dc39692098aa5
Branches Tags
retoor:main

No commits in common. "94e5d7bfa5d8444a4d3eb9772bf52ce0e60344ed" and "8e4a8f6f09fd766a320aaa49659dc39692098aa5" have entirely different histories.
94e5d7bfa5 ... 8e4a8f6f09

6 changed files with 37 additions and 100 deletions
Show Stats Expand all files Collapse all files

1
.gitignore vendored
View File

@ -1,3 +1,2 @@
.history .history
.vscode .vscode
publish

32
README.md
View File

@ -22,29 +22,15 @@ cat ./spam/example_spam1.txt | ./isspam
``` ```
## Example output ## Example output
``` ```
File: ./spam/example_spam3.txt File: ./not_spam/not_spam1.txt
Capitalized words: 39 Capitalized words: 1
Sentences: 20 Sentences: 5
Words: 420 Words: 52
Numbers: 1 Numbers: 0
Forbidden words: 15 Forbidden words: 0
<0:recovery> Word count per sentence: 10
<1:techie>
<2:https> Memory usage: 29 KB, 479 (re)allocated, 327 unqiue free'd, 0 in use.
<3:digital>
<4:hack>
<5://>
<6:com>
<7:@>
<8:crypto>
<9:bitcoin>
<10:whatsapp>
<11:cryptocurrency>
<12:stolen>
<13:contact>
<14:understanding>
Word count per sentence: 21
Memory usage: 1 MB, 6.460 (re)allocated, 4.222 unqiue free'd, 0 in use.
``` ```
## Valgrind status ## Valgrind status
Date: 2024-11-28 Date: 2024-11-28

BIN
isspam
View File

Binary file not shown.

84
isspam.c
View File

@ -24,28 +24,9 @@ char *forbidden_words[] = {
"stolen", "freeze", "quick", "crucial", "tracing", "scammers", "expers", "hire", "century", "stolen", "freeze", "quick", "crucial", "tracing", "scammers", "expers", "hire", "century",
"transaction", "essential", "managing", "contact", "contacting", "understanding", "assets", "funds", NULL}; "transaction", "essential", "managing", "contact", "contacting", "understanding", "assets", "funds", NULL};
bool show_capitalized = false;
bool show_sentences = false;
bool show_words = false;
bool show_numbers = false;
bool show_forbidden_words = true;
bool file_exists(char * path){
FILE * f = fopen(path, "r");
bool result = f != NULL;
if(f){
fclose(f);
}
return result;
}
void sld(sl *lst) { void sld(sl *lst) {
for (ulonglong i = 0; i < lst->count; i++) { for (uint i = 0; i < lst->count; i++) {
printf("<%llu:%s>\n", i, lst->strings[i]); printf("<%u:%s>\n", i, lst->strings[i]);
} }
} }
@ -68,10 +49,10 @@ char *remove_preserved_chars(char *content) {
char *slds(sl *lst) { char *slds(sl *lst) {
str_t *buffer = strn(1337); str_t *buffer = strn(1337);
for (ulonglong i = 0; i < lst->count; i++) { for (uint i = 0; i < lst->count; i++) {
char *temp = (char *)malloc(strlen(lst->strings[i]) + 20); char *temp = (char *)malloc(strlen(lst->strings[i]) + 20);
char *cc = remove_preserved_chars(lst->strings[i]); char *cc = remove_preserved_chars(lst->strings[i]);
sprintf(temp, "<%llu:%s>\n", i, cc); sprintf(temp, "<%u:%s>\n", i, cc);
free(cc); free(cc);
stra(buffer, temp); stra(buffer, temp);
free(temp); free(temp);
@ -263,6 +244,7 @@ sl *get_forbidden_words(char *content) {
slf(words); slf(words);
return found; return found;
} }
void analyze(FILE *f) { void analyze(FILE *f) {
char *data = fread_till_eof(f); char *data = fread_till_eof(f);
@ -275,10 +257,8 @@ void analyze(FILE *f) {
// All capitalized words // All capitalized words
sl *capitalized_words = get_capitalized_words(data); sl *capitalized_words = get_capitalized_words(data);
ulonglong capitalized_words_count = capitalized_words->count; uint capitalized_words_count = capitalized_words->count;
printf("Capitalized words: %llu\n", capitalized_words_count); printf("Capitalized words: %u\n", capitalized_words_count);
if(show_capitalized)
sld(capitalized_words);
sbuf = slds(capitalized_words); sbuf = slds(capitalized_words);
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
@ -286,9 +266,8 @@ void analyze(FILE *f) {
sl *sentences = get_sentences(data); sl *sentences = get_sentences(data);
// All sentences // All sentences
printf("Sentences: %llu\n", sentences->count); printf("Sentences: %u\n", sentences->count);
if(show_sentences) // sld(sentences);
sld(sentences);
sbuf = slds(sentences); sbuf = slds(sentences);
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
@ -296,33 +275,31 @@ void analyze(FILE *f) {
sl *words = get_words(data); sl *words = get_words(data);
// All words // All words
printf("Words: %llu\n", words->count); printf("Words: %u\n", words->count);
if(show_words) // sld(words);
sld(words);
sbuf = slds(words); sbuf = slds(words);
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
// Numbers // Numbers
sl *numbers = get_numbers(data); sl *numbers = get_numbers(data);
printf("Numbers: %llu\n", numbers->count); printf("Numbers: %u\n", numbers->count);
if(show_numbers) // sld(numbers);
sld(numbers);
sbuf = slds(numbers); sbuf = slds(numbers);
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
// Forbidden words // Forbidden words
sl *fw = get_forbidden_words(data); sl *fw = get_forbidden_words(data);
printf("Forbidden words: %llu\n", fw->count); printf("Forbidden words: %u\n", fw->count);
if(show_forbidden_words) // sld(fw);
sld(fw);
sbuf = slds(fw); sbuf = slds(fw);
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
strd(all); strd(all);
ulonglong word_count_per_sentence = words->count / (sentences->count ? sentences->count : 1); uint word_count_per_sentence = words->count / sentences->count;
printf("Word count per sentence: %llu\n", word_count_per_sentence); printf("Word count per sentence: %u\n", word_count_per_sentence);
slf(capitalized_words); slf(capitalized_words);
slf(sentences); slf(sentences);
@ -343,34 +320,11 @@ int main(int argc, char *argv[]) {
if (argc > 1) { if (argc > 1) {
for (int i = 1; i < argc; i++) { for (int i = 1; i < argc; i++) {
if(!strcmp(argv[1],"--hide-capitalized")){
show_capitalized=false;
}else if(!strcmp(argv[1],"--show-sentences")){
show_sentences=true;
}else if(!strcmp(argv[1],"--show-words")){
show_words=true;
}else if(!strcmp(argv[1],"--show-numbers")){
show_words=true;
}else if(!strcmp(argv[1],"--hide-forbidden-words")){
show_forbidden_words=false;
}else if(!strcmp(argv[1],"help") || !strcmp(argv[1],"--help")){
printf("%s",
"Usage: spam [file] [file] [file]\n"
"Flag defaults:\n"
" hide-capitalized = true\n"
" show-sentences = false\n"
" show-words = false\n"
" show-numbers = false\n"
" hide-forbidden-words = false\n");
return 0;
}
printf("File: %s\n", argv[i]); printf("File: %s\n", argv[i]);
analyze_file(argv[i]); analyze_file(argv[i]);
printf("%s\n", rmalloc_stats());
printf("\n"); printf("\n");
} }
printf("%s\n", rmalloc_stats());
return 0; return 0;
} }
analyze(stdin); analyze(stdin);

4
rmalloc.h
View File

@ -185,7 +185,7 @@ void *rfree(void *obj) {
char *rmalloc_lld_format(ulonglong num) { char *rmalloc_lld_format(ulonglong num) {
char res[100]; char res[100];
res[0] = 0; res[0] = 0;
sprintf(res, "%'llu", num); sprintf(res, "%'lld", num);
char *resp = res; char *resp = res;
while (*resp) { while (*resp) {
if (*resp == ',') if (*resp == ',')
@ -208,7 +208,7 @@ char *rmalloc_bytes_format(int factor, ulonglong num) {
} }
char *rmalloc_stats() { char *rmalloc_stats() {
static char res[300]; static char res[200];
res[0] = 0; res[0] = 0;
setlocale(LC_NUMERIC, "en_US.UTF-8"); setlocale(LC_NUMERIC, "en_US.UTF-8");
sprintf(res, "Memory usage: %s, %s (re)allocated, %s unqiue free'd, %s in use.", rmalloc_bytes_format(0, rmalloc_total_bytes_allocated), sprintf(res, "Memory usage: %s, %s (re)allocated, %s unqiue free'd, %s in use.", rmalloc_bytes_format(0, rmalloc_total_bytes_allocated),

10
rstring_list.h
View File

@ -4,11 +4,9 @@
#include <stdlib.h> #include <stdlib.h>
#include <string.h> #include <string.h>
typedef struct rstring_list_t { typedef struct rstring_list_t {
unsigned long long size; unsigned int size;
unsigned long long count; unsigned int count;
char **strings; char **strings;
} rstring_list_t; } rstring_list_t;
@ -19,7 +17,7 @@ rstring_list_t *rstring_list_new() {
} }
void rstring_list_free(rstring_list_t *rsl) { void rstring_list_free(rstring_list_t *rsl) {
for (unsigned long long i = 0; i < rsl->size; i++) { for (uint i = 0; i < rsl->size; i++) {
free(rsl->strings[i]); free(rsl->strings[i]);
} }
if (rsl->strings) if (rsl->strings)
@ -36,7 +34,7 @@ void rstring_list_add(rstring_list_t *rsl, char *str) {
rsl->count++; rsl->count++;
} }
bool rstring_list_contains(rstring_list_t *rsl, char *str) { bool rstring_list_contains(rstring_list_t *rsl, char *str) {
for (unsigned long long i = 0; i < rsl->count; i++) { for (uint i = 0; i < rsl->count; i++) {
if (!strcmp(rsl->strings[i], str)) if (!strcmp(rsl->strings[i], str))
return true; return true;
} }