works for the most part
This commit is contained in:
anon
parent
a4247754a5
commit
9fba401f78
1
.gitignore
vendored
1
.gitignore
vendored
@ -1,3 +1,4 @@
|
||||
*.out
|
||||
*.o
|
||||
*.yy.*
|
||||
.gdb_history
|
||||
|
||||
30
Makefile
30
Makefile
@ -1,17 +1,35 @@
|
||||
#LFLAGS := $(if $DEBUG, --debug --trace)
|
||||
LDLIBS := -lboost_string_algo
|
||||
.PHONY: clean test run
|
||||
|
||||
ifeq ($(DEBUG), 1)
|
||||
LFLAGS += --debug --trace
|
||||
CXXFLAGS += -Wall -Wextra -Wpedantic
|
||||
CXXFLAGS += -DDEBUG -O0 -ggdb -pg -fno-inline
|
||||
endif
|
||||
|
||||
OUT := a.out
|
||||
|
||||
SOURCE.d := source/
|
||||
OBJECT.d := object/
|
||||
TEST.d := test/
|
||||
|
||||
main:
|
||||
${LEX} ${LFLAGS} -o ${SOURCE.d}/main.yy.cpp ${SOURCE.d}/main.l
|
||||
${CXX} ${SOURCE.d}/main.yy.cpp -o ${OUT}
|
||||
SOURCE := main.cpp xml.cpp cli.cpp
|
||||
OBJECT := $(addprefix ${OBJECT.d}/,${SOURCE})
|
||||
OBJECT := ${OBJECT:.cpp=.o}
|
||||
OBJECT := ${OBJECT:.c=.o}
|
||||
|
||||
%.cpp: %.l
|
||||
${LEX} ${LFLAGS} -o $@ $<
|
||||
|
||||
${OBJECT.d}/%.o: ${SOURCE.d}/%.cpp
|
||||
${COMPILE.cpp} -o $@ $<
|
||||
|
||||
${OUT}: ${OBJECT}
|
||||
${LINK.cpp} -o $@ ${OBJECT} ${LDLIBS}
|
||||
|
||||
run:
|
||||
./${OUT} ${TEST.d}/draft.csml
|
||||
# ./${OUT} ${TEST.d}/draft.csml
|
||||
./${OUT} ${TEST.d}/draft.html
|
||||
|
||||
clean:
|
||||
-rm ${OUT}
|
||||
-rm ${OBJECT}
|
||||
|
||||
@ -44,19 +44,28 @@ and seemlessly translate it in the end.
|
||||
```
|
||||
<tag> [(<head>)] {<body>}
|
||||
```
|
||||
```
|
||||
<tag>;
|
||||
```
|
||||
Example:
|
||||
```
|
||||
div (?! find a stand alone attribute) { lorem ipsum }
|
||||
textarea (readonly) { lorem ipsum }
|
||||
```
|
||||
```
|
||||
br;
|
||||
```
|
||||
|
||||
The last __identifier__,
|
||||
defined by this regular expression: `[A-z][A-z0-9]*`,
|
||||
before a (optional) _head_ or _body_
|
||||
before a (optional) _head_, _body_ or semi-colon
|
||||
is considered to be a __tag__.
|
||||
CSML by itself does not enforce any (sub)set of words to be "valid"
|
||||
(however, related tools might).
|
||||
Each _tag_ is pushed into a stack and later popped by the end of a body being found.
|
||||
|
||||
If the _tag_ is followed by a semi-colon (';'),
|
||||
it's a self-closing tag.
|
||||
|
||||
The __head__ holds _attributes_.
|
||||
A missing _head_ signals that there are no attributes to be translated.
|
||||
Any text may be a valid __attribute__.
|
||||
@ -78,6 +87,7 @@ List of escaped special characters:
|
||||
+ \}
|
||||
+ \,
|
||||
+ \:
|
||||
+ \;
|
||||
Note, that they are not requred to be always escaped,
|
||||
but are highly advised.
|
||||
|
||||
@ -118,7 +128,7 @@ are different sides of the same coin.
|
||||
```
|
||||
?! [options] <file>+
|
||||
-c : the input is to be force interpeted as CSML
|
||||
-? : the input is to be force interpeted as XML/HTML
|
||||
-x : the input is to be force interpeted as XML/HTML
|
||||
-o <file> : specify output file name for the NEXT file
|
||||
-q <char> : use <char> for quoting (default: "'")
|
||||
-v : print version and quit
|
||||
|
||||
0
object/.gitkeep
Normal file
0
object/.gitkeep
Normal file
46
source/cli.cpp
Normal file
46
source/cli.cpp
Normal file
@ -0,0 +1,46 @@
|
||||
#include "cli.h"
|
||||
|
||||
#include <getopt.h>
|
||||
#include <stdio.h>
|
||||
|
||||
const char * const version =
|
||||
# include "version.inc"
|
||||
;
|
||||
|
||||
const char * help_message =
|
||||
"?! [options] <file>+\n"
|
||||
" -c : the input is to be force interpeted as CSML\n"
|
||||
" -x : the input is to be force interpeted as XML/HTML\n"
|
||||
" -o <file> : specify output file name for the NEXT file\n"
|
||||
" -q <char> : use <char> for quoting (default: \"'\")\n"
|
||||
" -v : print version and quit\n"
|
||||
" -h : print help and quit\n"
|
||||
;
|
||||
|
||||
signed parse_r1_arguments(int argc, char * * argv){
|
||||
const char * const optstring = "hv" "cxo:q:";
|
||||
|
||||
static struct option long_options[] = {
|
||||
{"help", no_argument, 0, 'h'},
|
||||
{"version", no_argument, 0, 'v'},
|
||||
{0, 0, 0, 0}
|
||||
};
|
||||
|
||||
char opt;
|
||||
while ((opt = getopt_long(argc, argv, optstring, long_options, NULL)) != -1) {
|
||||
switch (opt) {
|
||||
case 'h': {
|
||||
fputs(help_message, stdout);
|
||||
} return 1;
|
||||
case 'v': {
|
||||
fputs(version, stdout);
|
||||
} return 1;
|
||||
case '?': {
|
||||
} return 2;
|
||||
default: {
|
||||
} break;
|
||||
}
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
9
source/cli.h
Normal file
9
source/cli.h
Normal file
@ -0,0 +1,9 @@
|
||||
#ifndef CLI_H
|
||||
|
||||
/* Parse arguments with perminant effects (-h)
|
||||
* Perform validation.
|
||||
*/
|
||||
signed parse_r1_arguments(int argc, char * * argv);
|
||||
|
||||
#define CLI_H
|
||||
#endif
|
||||
@ -1,12 +1,11 @@
|
||||
%{
|
||||
#include "scanner.h"
|
||||
|
||||
#include <stdio.h>
|
||||
#include <stack>
|
||||
#include <string>
|
||||
#include <boost/algorithm/string/trim.hpp>
|
||||
|
||||
#define ECHOS(s) fwrite(s, strlen(s), sizeof(char), yyout)
|
||||
#define ECHO_CANDIDATE _ECHO_CANDIDATE()
|
||||
|
||||
std::stack<std::string> tag_stack;
|
||||
std::string tag_candidate = "";
|
||||
|
||||
@ -77,6 +76,10 @@ identifier [A-z][A-z0-9]*
|
||||
ECHOS(("<" + tag_stack.top() + " ").c_str());
|
||||
BEGIN HEAD;
|
||||
}
|
||||
; {
|
||||
ECHOS(("<" + tag_candidate + "/>").c_str());
|
||||
tag_candidate = "";
|
||||
}
|
||||
\{ {
|
||||
push_tag();
|
||||
ECHOS(("<" + tag_stack.top() + ">").c_str());
|
||||
@ -156,8 +159,3 @@ identifier [A-z][A-z0-9]*
|
||||
}
|
||||
|
||||
%%
|
||||
|
||||
signed main(int argc, char * * argv) {
|
||||
yyin = fopen(argv[1], "r");
|
||||
yylex();
|
||||
}
|
||||
108
source/main.cpp
Normal file
108
source/main.cpp
Normal file
@ -0,0 +1,108 @@
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include <stddef.h>
|
||||
#include <libgen.h>
|
||||
|
||||
#include "cli.h"
|
||||
|
||||
extern FILE * yyin;
|
||||
extern FILE * yyout;
|
||||
extern int yylex(void);
|
||||
|
||||
const char DEFAULT_QUOTE = '\'';
|
||||
char quote = DEFAULT_QUOTE;
|
||||
char * output = NULL;
|
||||
char * input = NULL;
|
||||
const char * const csml_extension = ".csml";
|
||||
|
||||
enum input_type_t {
|
||||
CSML,
|
||||
XML,
|
||||
AUTO_DETECT,
|
||||
} input_type = AUTO_DETECT;
|
||||
|
||||
const char * const auto_output_extensions[] = {csml_extension, ".html"};
|
||||
|
||||
char * output_name_from_input_name(const char * const input, const char * const extension) {
|
||||
char * input_duplicate = strdup(input);
|
||||
char * dn = strdup(dirname(input_duplicate));
|
||||
char * bn = strdup(basename(input_duplicate));
|
||||
free(input_duplicate);
|
||||
|
||||
char * r = (char *)malloc(strlen(dn) + strlen(bn) + strlen(extension));
|
||||
strcpy(r, dn);
|
||||
strcat(r, bn);
|
||||
|
||||
const char *last_dot = strrchr(bn, '.');
|
||||
if (last_dot) {
|
||||
strncpy(r, bn, last_dot - bn);
|
||||
} else {
|
||||
strcpy(r, bn);
|
||||
}
|
||||
|
||||
strcat(r, extension);
|
||||
|
||||
free(dn);
|
||||
free(bn);
|
||||
|
||||
return r;
|
||||
}
|
||||
|
||||
signed main(int argc, char * * argv) {
|
||||
switch (parse_r1_arguments(argc - 1, argv + 1)) {
|
||||
case 1: {
|
||||
} return 0;
|
||||
case 2: {
|
||||
} return 1;
|
||||
default: {
|
||||
} break;
|
||||
}
|
||||
|
||||
for (int n = argc - 1; n; n--) {
|
||||
if (!strcmp(argv[n], "-c")) {
|
||||
input_type = input_type_t::CSML;
|
||||
} else if (!strcmp(argv[n], "-x")) {
|
||||
input_type = input_type_t::XML;
|
||||
} else if (!strcmp(argv[n], "-q")) {
|
||||
++n;
|
||||
quote = argv[n][0];
|
||||
} else if (!strcmp(argv[n], "-o")) {
|
||||
++n;
|
||||
output = argv[n];
|
||||
} else {
|
||||
/* --- Preparation --- */
|
||||
input = argv[n];
|
||||
|
||||
if (input_type == AUTO_DETECT) {
|
||||
if (!strcmp(input + strlen(input) - (sizeof(csml_extension)-1), csml_extension)) {
|
||||
input_type = CSML;
|
||||
} else {
|
||||
input_type = XML;
|
||||
}
|
||||
}
|
||||
|
||||
if (output) {
|
||||
yyout = fopen(output, "w");
|
||||
} else {
|
||||
yyout = stdout;
|
||||
}
|
||||
|
||||
yyin = fopen(input, "r");
|
||||
yyout = fopen(output, "w");
|
||||
|
||||
/* --- Meat --- */
|
||||
yylex();
|
||||
|
||||
/* --- Clean up --- */
|
||||
if (yyin != stdin) {
|
||||
fclose(yyin);
|
||||
}
|
||||
if (yyout != stdout) {
|
||||
fclose(yyout);
|
||||
}
|
||||
free(output);
|
||||
output = NULL;
|
||||
}
|
||||
}
|
||||
}
|
||||
6
source/scanner.h
Normal file
6
source/scanner.h
Normal file
@ -0,0 +1,6 @@
|
||||
#ifndef SCANNER_H
|
||||
|
||||
#define ECHOS(s) fwrite(s, strlen(s), sizeof(char), yyout)
|
||||
|
||||
#define SCANNER_H
|
||||
#endif
|
||||
1
source/version.inc
Normal file
1
source/version.inc
Normal file
@ -0,0 +1 @@
|
||||
"v0.4"
|
||||
115
source/xml.l
Normal file
115
source/xml.l
Normal file
@ -0,0 +1,115 @@
|
||||
%{
|
||||
#include "scanner.h"
|
||||
|
||||
#include <stdio.h>
|
||||
|
||||
bool is_comment_multiline = false;
|
||||
unsigned long long comment_begining;
|
||||
|
||||
%}
|
||||
|
||||
%option noyywrap
|
||||
%option nodefault
|
||||
|
||||
%x TAG_START TAG_MAYBE TAG COMMENT STRING
|
||||
|
||||
ws [ \t\r\v\f]
|
||||
wsnl [ \t\r\v\f\n]
|
||||
identifier [A-z][A-z0-9]*
|
||||
|
||||
%%
|
||||
<INITIAL>{
|
||||
\< {
|
||||
BEGIN TAG_START;
|
||||
}
|
||||
\<\!-- {
|
||||
comment_begining = ftell(yyin);
|
||||
ECHOS("//");
|
||||
BEGIN COMMENT;
|
||||
}
|
||||
.|\n {
|
||||
ECHO;
|
||||
}
|
||||
}
|
||||
|
||||
<COMMENT>{
|
||||
. {
|
||||
ECHO;
|
||||
}
|
||||
\n {
|
||||
is_comment_multiline = true;
|
||||
}
|
||||
--\> {
|
||||
if (is_comment_multiline) {
|
||||
auto buffer = ftell(yyin);
|
||||
fseek(yyin, comment_begining+1, SEEK_SET);
|
||||
fputc('*', yyin);
|
||||
fseek(yyin, buffer, SEEK_SET);
|
||||
}
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
}
|
||||
|
||||
<TAG_START>{
|
||||
\/{identifier}+{wsnl}*\> {
|
||||
ECHOS("}");
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
{identifier}+ {
|
||||
ECHO;
|
||||
BEGIN TAG_MAYBE;
|
||||
}
|
||||
}
|
||||
|
||||
<TAG_MAYBE>{
|
||||
\> {
|
||||
ECHOS(" {");
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
\/\> {
|
||||
ECHOS(";");
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
{wsnl} {
|
||||
ECHO;
|
||||
}
|
||||
. {
|
||||
yyless(0);
|
||||
ECHOS("(");
|
||||
BEGIN TAG;
|
||||
}
|
||||
}
|
||||
|
||||
<TAG>{
|
||||
\"|\' {
|
||||
BEGIN STRING;
|
||||
}
|
||||
= {
|
||||
ECHOS(": ");
|
||||
}
|
||||
\> {
|
||||
ECHOS(") {");
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
\/\> {
|
||||
ECHOS(";");
|
||||
BEGIN INITIAL;
|
||||
}
|
||||
.|\n {
|
||||
ECHO;
|
||||
}
|
||||
}
|
||||
|
||||
<STRING>{
|
||||
[^\\]\"|\' {
|
||||
BEGIN TAG;
|
||||
}
|
||||
, {
|
||||
ECHOS("\\,");
|
||||
}
|
||||
.|\n {
|
||||
ECHO;
|
||||
}
|
||||
}
|
||||
|
||||
%%
|
||||
@ -1,9 +1,10 @@
|
||||
// DOCTYPE HTML
|
||||
// DOCTYPE HTML
|
||||
html {
|
||||
head {
|
||||
}
|
||||
body {
|
||||
div(class: myclass) {
|
||||
hr;
|
||||
div (class: myclass) {
|
||||
lorem ipsum
|
||||
}
|
||||
}
|
||||
|
||||
12
test/draft.html
Normal file
12
test/draft.html
Normal file
@ -0,0 +1,12 @@
|
||||
./a.out test//draft.csml
|
||||
<!-- DOCTYPE HTML -->
|
||||
<html>
|
||||
<head>
|
||||
</head>
|
||||
<body>
|
||||
<hr/>
|
||||
<div class='myclass'>
|
||||
lorem ipsum
|
||||
</div>
|
||||
</body>
|
||||
</html>
|
||||
Loading…
x
Reference in New Issue
Block a user