summaryrefslogtreecommitdiff
path: root/src/mcpp-source.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/mcpp-source.c')
-rw-r--r--src/mcpp-source.c332
1 files changed, 332 insertions, 0 deletions
diff --git a/src/mcpp-source.c b/src/mcpp-source.c
new file mode 100644
index 0000000..c40e370
--- /dev/null
+++ b/src/mcpp-source.c
@@ -0,0 +1,332 @@
+#include <defs.h>
+
+static char *
+read_stream_bytes( FILE *fp, size_t *file_length )
+{
+ char *data = NULL;
+ size_t length = 0;
+ size_t capacity = 0;
+
+ if( fp == NULL || file_length == NULL )
+ {
+ errno = EINVAL;
+ return( NULL );
+ }
+
+ *file_length = 0;
+
+ for( ;; )
+ {
+ size_t n;
+
+ if( capacity - length < 4096 )
+ {
+ size_t new_capacity = capacity ? capacity * 2 : 8192;
+ char *q = (char *)realloc( data, new_capacity + 1 );
+ if( q == NULL )
+ {
+ free( data );
+ return( NULL );
+ }
+ data = q;
+ capacity = new_capacity;
+ }
+
+ n = fread( data + length, 1, capacity - length, fp );
+ length += n;
+
+ if( n == 0 )
+ {
+ if( ferror(fp) )
+ {
+ free( data );
+ return( NULL );
+ }
+ break;
+ }
+ }
+
+ if( data == NULL )
+ {
+ data = (char *)calloc( 1, 1 );
+ if( data == NULL )
+ return( NULL );
+ }
+ else
+ data[length] = 0;
+
+ *file_length = length;
+ return( data );
+}
+
+static int
+normalize_newlines( mcpu_text *text )
+{
+ size_t r;
+ size_t w = 0;
+
+ if( text == NULL || text->data == NULL )
+ return( 0 );
+
+ for( r = 0; r < text->length; ++r )
+ {
+ if( text->data[r] == '\r' )
+ {
+ if( r + 1 < text->length && text->data[r + 1] == '\n' )
+ ++r;
+ text->data[w++] = '\n';
+ }
+ else
+ text->data[w++] = text->data[r];
+ }
+
+ text->length = w;
+ text->data[w] = 0;
+
+ return( 0 );
+}
+
+void
+mcpu_source_init( mcpu_source *source )
+{
+ if( source == NULL )
+ return;
+
+ source->filename = NULL;
+ mcpu_text_init( &source->text );
+ mcpu_text_init( &source->logical_line );
+ source->splice_offsets = NULL;
+ source->splice_count = 0;
+ source->splice_capacity = 0;
+ source->offset = 0;
+ source->line = 1;
+}
+
+void
+mcpu_source_free( mcpu_source *source )
+{
+ if( source == NULL )
+ return;
+
+ free( source->filename );
+ source->filename = NULL;
+ mcpu_text_free( &source->text );
+ mcpu_text_free( &source->logical_line );
+ free( source->splice_offsets );
+ source->splice_offsets = NULL;
+ source->splice_count = 0;
+ source->splice_capacity = 0;
+ source->offset = 0;
+ source->line = 1;
+}
+
+static int
+source_from_bytes( mcpu_source *source, char *bytes, size_t byte_length,
+ const char *name )
+{
+ const __mpu_char8_t *input = (const __mpu_char8_t *)bytes;
+ const __mpu_char8_t *p;
+
+ if( memchr( bytes, 0, byte_length ) != NULL )
+ {
+ fprintf( stderr, "%s: NUL character is not allowed in source text\n", name );
+ return( -1 );
+ }
+
+ if( byte_length >= 3 &&
+ (unsigned char)bytes[0] == 0xef &&
+ (unsigned char)bytes[1] == 0xbb &&
+ (unsigned char)bytes[2] == 0xbf )
+ input += 3;
+
+ if( !mpu_utf8valid(input) )
+ {
+ fprintf( stderr, "%s: invalid UTF-8\n", name );
+ return( -1 );
+ }
+
+ p = input;
+ while( *p != 0 )
+ {
+ __mpu_char32_t value;
+ const __mpu_char8_t *next;
+ __mpu_char16_t c;
+
+ next = mpu_utf8get( p, &value );
+ if( next == NULL )
+ {
+ fprintf( stderr, "%s: invalid UTF-8\n", name );
+ return( -1 );
+ }
+
+ if( value <= 0xffff )
+ c = (__mpu_char16_t)value;
+ else
+ c = MCPU_CPP_NON_UCS2_SENTINEL;
+
+ if( mcpu_text_append_char(&source->text, c) != 0 )
+ return( -1 );
+
+ p = next;
+ }
+
+ if( normalize_newlines(&source->text) != 0 )
+ return( -1 );
+
+ source->filename = strdup( name );
+ if( source->filename == NULL )
+ return( -1 );
+
+ return( 0 );
+}
+
+int
+mcpu_source_valid_ucs2( const __mpu_char16_t *text, size_t length )
+{
+ size_t i;
+
+ if( text == NULL && length != 0 )
+ return( 0 );
+
+ for( i = 0; i < length; ++i )
+ if( text[i] == MCPU_CPP_NON_UCS2_SENTINEL )
+ return( 0 );
+
+ return( 1 );
+}
+
+int
+mcpu_source_open_stream( mcpu_source *source, FILE *stream, const char *name )
+{
+ char *bytes;
+ size_t byte_length;
+ int rc;
+
+ if( source == NULL || stream == NULL || name == NULL )
+ {
+ errno = EINVAL;
+ return( -1 );
+ }
+
+ mcpu_source_free( source );
+ mcpu_source_init( source );
+
+ bytes = read_stream_bytes( stream, &byte_length );
+ if( bytes == NULL )
+ return( -1 );
+
+ rc = source_from_bytes( source, bytes, byte_length, name );
+ free( bytes );
+ return( rc );
+}
+
+int
+mcpu_source_open( mcpu_source *source, const char *filename )
+{
+ FILE *fp;
+ int rc;
+
+ if( source == NULL || filename == NULL )
+ {
+ errno = EINVAL;
+ return( -1 );
+ }
+
+ fp = fopen( filename, "rb" );
+ if( fp == NULL )
+ return( -1 );
+
+ rc = mcpu_source_open_stream( source, fp, filename );
+ if( fclose(fp) != 0 && rc == 0 )
+ rc = -1;
+
+ return( rc );
+}
+
+static int
+append_splice_offset( mcpu_source *source, size_t offset )
+{
+ size_t *offsets;
+ size_t capacity;
+
+ if( source->splice_count < source->splice_capacity )
+ {
+ source->splice_offsets[source->splice_count++] = offset;
+ return( 0 );
+ }
+
+ capacity = source->splice_capacity ? source->splice_capacity * 2 : 8;
+ offsets = (size_t *)realloc( source->splice_offsets,
+ capacity * sizeof(*offsets) );
+ if( offsets == NULL )
+ return( -1 );
+
+ source->splice_offsets = offsets;
+ source->splice_capacity = capacity;
+ source->splice_offsets[source->splice_count++] = offset;
+ return( 0 );
+}
+
+
+int
+mcpu_source_next_line( mcpu_source *source,
+ const __mpu_char16_t **line,
+ size_t *length,
+ unsigned *line_number,
+ unsigned *next_line_number,
+ int *spliced )
+{
+ unsigned first_line;
+
+ if( source == NULL || line == NULL || length == NULL ||
+ line_number == NULL || next_line_number == NULL || spliced == NULL )
+ {
+ errno = EINVAL;
+ return( -1 );
+ }
+
+ if( source->offset >= source->text.length )
+ return( 0 );
+
+ mcpu_text_free( &source->logical_line );
+ mcpu_text_init( &source->logical_line );
+ source->splice_count = 0;
+ first_line = source->line;
+ *spliced = 0;
+
+ while( source->offset < source->text.length )
+ {
+ __mpu_char16_t c = source->text.data[source->offset++];
+
+ /*
+ * Backslash-newline deletion is preprocessing phase 2 and therefore
+ * happens before comments, directives and macro recognition.
+ */
+ if( c == '\\' && source->offset < source->text.length &&
+ source->text.data[source->offset] == '\n' )
+ {
+ if( append_splice_offset(source, source->logical_line.length) != 0 )
+ return( -1 );
+ ++source->offset;
+ ++source->line;
+ *spliced = 1;
+ continue;
+ }
+
+ if( mcpu_text_append_char( &source->logical_line, c ) != 0 )
+ return( -1 );
+
+ if( c == '\n' )
+ {
+ ++source->line;
+ break;
+ }
+ }
+
+ *line = source->logical_line.data;
+ *length = source->logical_line.length;
+ *line_number = first_line;
+ *next_line_number = source->line;
+
+ return( 1 );
+}
generated by cgit v1.2.3 (git 2.39.0) at 2026-10-03 17:15:22 +0300