static void skip_to_endln(state_t* state)
{
char c;
while (ppimpl_has_input(state))
{
c = ppimpl_get_input(state);
if (c == '\n')
return;
}
}
///
/// Performs preprocessing using the internally
/// constructed state object.
///
void ppimpl_process(state_t* state)
{
char c;
match_t* m;
bool reprocess = false;
size_t i;
scope_t* current;
while (ppimpl_has_input(state))
{
// Fetch a character from the input.
c = ppimpl_get_input(state);
// Push it onto the output.
list_append(&state->cached_output, (void*)c);
// Update string tracking states.
ppimpl_track_strings(state);
// If we are in a single or double string, no matching
// operations can be performed.
if (state->in_single_string || state->in_double_string)
continue;
// We keep process once unless we're asked to restart
// by a match. This is used when a match changes the
// handlers (such as appending a new handler or removing
// an old one), and we need to restart the matching process.
do
{
reprocess = false;
// Iterate through all of our match handlers.
for (i = 0; i < list_size(&state->handlers); i++)
{
m = (match_t*)list_get_at(&state->handlers, i);
// Test for a match and if so, handle it.
if (ppimpl_match_test(state, m))
{
// Remove the characters from the cached input
// depending on the match text length.
list_delete_range(&state->cached_output,
list_size(&state->cached_output) - blength(m->text.ref),
list_size(&state->cached_output) - 1);
// Reset the printing index.
state->print_index = 0;
// Call the match handler.
m->handler(state, m, &reprocess);
if (reprocess)
break;
}
}
}
while (reprocess);
}
// We have finished processing; everything is stored in the cached output,
// so now we push everything that was in the cached output.
while (list_size(&state->cached_output) > 0)
state->output((char)list_extract_at(&state->cached_output, 0));
}
///
/// Checks to see if the current output cache
/// has a match against a specified match definition.
///
bool ppimpl_match_test(state_t* state, match_t* match)
{
size_t i;
int32_t j;
// Check if the output cache is even long enough to hold
// the match.
if (list_size(&state->cached_output) < (size_t)blength(match->text.ref))
return false;
// Check each character until we find one that doesn't
// match, going backwards through the list (remember that
// we're matching against the end of the list).
for (i = 0; i < (size_t)blength(match->text.ref); i++)
{
if (!match->case_insensitive)
{
if ((char)list_get_at(&state->cached_output, list_size(&state->cached_output) - i - 1) !=
(char)match->text.ref->data[blength(match->text.ref) - i - 1])
return false;
}
else
{
if (tolower((char)list_get_at(&state->cached_output, list_size(&state->cached_output) - i - 1)) !=
tolower((char)match->text.ref->data[blength(match->text.ref) - i - 1]))
return false;
}
}
// We have a match, ensure that it is at the start of the line
// if required.
if (match->line_start_only)
{
// We have to make sure that we only have spaces or tabs between
// the first character of the text and a \n or the start of the output.
j = (int32_t)(list_size(&state->cached_output) - (size_t)blength(match->text.ref) - 1);
while (j >= 0)
{
switch ((char)list_get_at(&state->cached_output, j))
{
case ' ':
case '\t':
// Permitted; continue.
break;
case '\n':
// This means that it is at the start of a line.
return true;
default:
// Anything else means that it's not at the start.
return false;
}
j--;
}
// If we reach here, check if i == 0, which means that we've
// matched the start of the input (which also counts as a match).
if (j <= 0)
return true;
else
assert(false); // Something is not right.
}
// We have a match, ensure it is a proper identifier.
if (match->identifier_only)
{
// We have to get the surrounding characters.
bool has_next = ppimpl_has_input(state);
bool has_previous = (list_size(&state->cached_output) >= (size_t)blength(match->text.ref) + 1);
char next;
char previous;
bool is_identifier = false;
if (has_next)
{
next = ppimpl_get_input(state);
list_insert_at(&state->cached_input, (void*)next, 0);
}
if (has_previous)
previous = (char)list_get_at(&state->cached_output, list_size(&state->cached_output) - blength(match->text.ref) - 1);
// Check to see if the preceeding character (before the
// match) and the proceeding character (after the match)
// both isolate this match as an identifier.
if (has_next && has_previous)
is_identifier = ppimpl_isolates(next, false) && ppimpl_isolates(previous, true);
else if (has_next)
is_identifier = ppimpl_isolates(next, false);
else if (has_previous)
is_identifier = ppimpl_isolates(previous, true);
else
is_identifier = true;
return is_identifier;
}
// We have a match that can be matched anywhere.
return true;
}
static void macro_handle(state_t* state, match_t* match, bool* reprocess)
{
bstring name;
bstring temp;
list_t parameters;
list_t arguments;
bool getting_name = true;
bool getting_parameters = false;
bool getting_arguments = false;
struct replace_info* info = match->userdata;
int i = 0;
int argument_brackets = 0;
char c;
char* start_loc;
match_t* new_match;
struct replace_info* new_info;
// Parse the parameters out of the name.
list_init(¶meters);
temp = bfromcstr("");
for (i = 0; i < blength(info->full); i++)
{
c = info->full->data[i];
if (getting_name)
{
if (c == '(')
{
getting_name = false;
getting_parameters = true;
name = bstrcpy(temp);
bassigncstr(temp, "");
}
else
bconchar(temp, c);
}
else if (getting_parameters)
{
if (c == ',' || c == ')')
{
btrimws(temp);
list_append(¶meters, bstrcpy(temp));
bassigncstr(temp, "");
if (c == ')')
{
getting_parameters = false;
break;
}
}
else
bconchar(temp, c);
}
}
// Attempt to accept an open bracket.
c = ppimpl_get_input(state);
while (c == '\1')
{
// Consume macro termination.
i = 0;
while (i < strlen("\1MACROTERMINATE\1"))
{
if (c != "\1MACROTERMINATE\1"[i++])
dhalt(ERR_PP_EXPECTED_OPEN_BRACKET, ppimpl_get_location(state));
c = ppimpl_get_input(state);
}
ppimpl_pop_scope(state);
}
if (c != '(')
dhalt(ERR_PP_EXPECTED_OPEN_BRACKET, ppimpl_get_location(state));
// Read arguments.
getting_arguments = true;
list_init(&arguments);
start_loc = ppimpl_get_location(state);
bassigncstr(temp, "");
while (ppimpl_has_input(state) && getting_arguments)
{
c = ppimpl_get_input(state);
if (c == '(')
{
argument_brackets++;
bconchar(temp, c);
}
else if (c == ')' && argument_brackets != 0)
{
argument_brackets--;
bconchar(temp, c);
}
else if (c == ')' && argument_brackets == 0)
{
list_append(&arguments, bstrcpy(temp));
bassigncstr(temp, "");
getting_arguments = false;
break;
}
else if (c == ',' && argument_brackets == 0)
{
list_append(&arguments, bstrcpy(temp));
bassigncstr(temp, "");
}
else
bconchar(temp, c);
}
if (getting_arguments)
dhalt(ERR_PP_NO_TERMINATING_BRACKET, start_loc);
// Check to see if the argument count is correct.
if (list_size(&arguments) > list_size(¶meters))
dhalt(ERR_PP_TOO_MANY_PARAMETERS, start_loc);
else if (list_size(&arguments) < list_size(¶meters))
dhalt(ERR_PP_NOT_ENOUGH_PARAMETERS, start_loc);
free(start_loc);
// Create a new scope for macro evaluation.
ppimpl_push_scope(state, true);
// Define the new handlers.
for (i = 0; i < list_size(¶meters); i++)
{
new_info = malloc(sizeof(struct replace_info));
new_info->full = list_get_at(¶meters, i);
new_info->replacement = list_get_at(&arguments, i);
if (biseq(new_info->full, new_info->replacement))
{
free(new_info);
continue;
}
new_match = malloc(sizeof(match_t));
new_match->text = bautofree(list_get_at(¶meters, i));
new_match->handler = replace_handle;
new_match->line_start_only = false;
new_match->identifier_only = true;
new_match->userdata = new_info;
new_match->case_insensitive = false;
ppimpl_register(state, new_match);
}
// Print out the macro evaluation and terminator.
ppimpl_printf(state, "%s\1MACROTERMINATE\1", info->replacement->data);
}
static void define_handle(state_t* state, match_t* match, bool* reprocess)
{
// We need to parse this manually because we're interested in getting
// the first word and then all of the content until a line that doesn't end
// with "\".
bstring name = bfromcstr("");
bstring word = bfromcstr("");
bstring definition = bfromcstr("");
bool getting_word = true;
bool getting_definition = true;
bool is_macro = false;
match_t* new_match;
struct replace_info* info;
// Get the first word.
while (getting_word)
{
char c = ppimpl_get_input(state);
bconchar(word, c);
if (!is_macro && c != '(')
bconchar(name, c);
bltrimws(word);
// Handle termination.
if (blength(word) > 0 && (c == ' ' || c == '\t') && !is_macro)
{
// End of word.
btrimws(word);
btrimws(name);
getting_word = false;
}
else if (blength(word) > 0 && c == '(' && !is_macro)
{
// Start of macro.
is_macro = true;
}
else if (blength(word) > 0 && c == '(' && is_macro)
{
// Second ( in a macro; error.
dhalt(ERR_PP_MACRO_MALFORMED, ppimpl_get_location(state));
}
else if (blength(word) > 0 && c == ')' && is_macro)
{
// End of macro name.
btrimws(word);
btrimws(name);
getting_word = false;
}
else if (blength(word) == 0 && c == '\n')
dhalt(ERR_PP_C_DEFINE_PARAMETERS_INCORRECT, ppimpl_get_location(state));
else if (blength(word) > 0 && c == '\n')
{
// End of word.
btrimws(word);
btrimws(name);
getting_word = false;
getting_definition = false;
ppimpl_printf(state, "\n");
}
}
// Get the definition.
while (getting_definition)
{
char c = ppimpl_get_input(state);
bconchar(definition, c);
bltrimws(definition);
if (c == '\n')
{
if (blength(definition) > 1 && definition->data[blength(definition) - 2] == '\\')
{
// Remove the new slash.
bdelete(definition, blength(definition) - 2, 1);
ppimpl_oprintf(state, "\n");
}
else
{
btrimws(definition);
getting_definition = false;
ppimpl_printf(state, "\n");
}
}
else if (c == '/' || c == '*')
{
if (blength(definition) > 1 && definition->data[blength(definition) - 2] == '/')
{
// a line or block comment
ppimpl_iprintf(state, "/%c", c);
// remove the slashes
bdelete(definition, blength(definition) - 2, 2);
btrimws(definition);
getting_definition = false;
}
}
}
if (blength(definition) == 0 && !is_macro)
bassigncstr(definition, "1");
// Create the new replacement handler.
info = malloc(sizeof(struct replace_info));
info->full = word;
info->replacement = definition;
if (biseq(info->full, info->replacement))
{
free(info);
return;
}
new_match = malloc(sizeof(match_t));
new_match->text = bautofree(name);
if (is_macro)
new_match->handler = macro_handle;
else
new_match->handler = replace_handle;
new_match->line_start_only = false;
new_match->identifier_only = true;
new_match->userdata = info;
new_match->case_insensitive = false;
ppimpl_register(state, new_match);
*reprocess = true;
}
static bstring skip_to_endif(state_t* state, bool stop_at_else, bool* stopped_at_else)
{
char c;
bool fresh_line = true;
bstring temp = bfromcstr("");
bstring temp_output = bfromcstr("");
bstring output = bfromcstr("");
int if_open = 1;
while (ppimpl_has_input(state))
{
c = ppimpl_get_input(state);
switch(c)
{
case '#':
case '.':
if (!fresh_line)
{
bconchar(output, c);
break;
}
bassigncstr(temp_output, "#");
// first skip spaces
while (ppimpl_has_input(state))
{
c = ppimpl_get_input(state);
bconchar(temp_output, c);
if (c != ' ' && c != '\t')
break;
}
// read pp directive
bassigncstr(temp, "");
bconchar(temp, c);
while (ppimpl_has_input(state))
{
c = ppimpl_get_input(state);
bconchar(temp_output, c);
if (c == ' ' || c == '\t' || c == '\n')
break;
bconchar(temp, c);
}
btolower(temp);
if (biseq(temp, bfromcstr("endif")))
{
if_open--;
if (if_open == 0)
{
if (c != '\n') skip_to_endln(state);
*stopped_at_else = false;
return output;
}
}
else if (biseq(temp, bfromcstr("if")))
{
if_open++;
}
else if (biseq(temp, bfromcstr("else")) && stop_at_else)
{
if (if_open == 1)
{
if (c != '\n') skip_to_endln(state);
*stopped_at_else = true;
return output;
}
}
bconcat(output, temp_output);
fresh_line = (c == '\n');
break;
case '\n':
fresh_line = true;
bconchar(output, c);
break;
case ' ':
case '\t':
bconchar(output, c);
break;
default:
fresh_line = false;
bconchar(output, c);
break;
}
}
// No .ENDIF was found.
dhalt(ERR_PP_ASM_NO_ENDIF_TO_IF, ppimpl_get_location(state));
// dhalt will trigger before this, but the compiler warns about
// control potentially reaching the end of this function.
return NULL;
}