6502/acme
1
0
Fork 0
mirror of https://github.com/uffejakobsen/acme.git synced 2025-02-10 11:31:49 +00:00
Code Issues Projects Releases Wiki Activity

now ignores BOM at start of source file

Browse Source 
git-svn-id: https://svn.code.sf.net/p/acme-crossass/code-0/trunk@420 4df02467-bbd4-4a76-a152-e7ce94205b78
This commit is contained in:
marcobaye 2024-09-05 09:47:49 +00:00
parent b7037ea990
commit ad9ef576ec
2 changed files with 36 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
src/input.c
View File

@ -20,7 +20,8 @@
// values for current_input.state
enum inputstate {
INPUTSTATE_SOF, // start of file (check for hashbang)
INPUTSTATE_SOF, // start of file (check for BOM and hashbang)
INPUTSTATE_BOMKLUGE, // send second byte of non-BOM and then the byte read from file
INPUTSTATE_NORMAL, // everything's fine
INPUTSTATE_AGAIN, // re-process last byte
INPUTSTATE_SKIPBLANKS, // shrink multiple spaces
@ -161,7 +162,7 @@ static void report_srcchar(int new_char)
}
// Deliver source code from current file (!) in shortened high-level format
// deliver source code from current file (!) in shortened high-level format
static char get_processed_from_file(void)
{
static int from_file = 0;
@ -171,8 +172,35 @@ static char get_processed_from_file(void)
case INPUTSTATE_SOF:
// fetch first byte from the current source file
from_file = getc(current_input.u.fd);
// check for bogus/malformed BOM (0xef 0xbb 0xbf as UTF-8-encoded 0xfeff)
if (from_file == 0xef) {
// first byte looks like BOM, so check second:
from_file = getc(current_input.u.fd);
if (from_file == 0xbb) {
// first two bytes look like BOM, so check third:
from_file = getc(current_input.u.fd);
if (from_file == 0xbf) {
// found BOM, so ignore
current_input.state = INPUTSTATE_NORMAL;
break;
} else {
// third byte does not match, so return first byte and make sure the others are delivered later
IF_WANTED_REPORT_SRCCHAR(0xef);
IF_WANTED_REPORT_SRCCHAR(0xbb);
IF_WANTED_REPORT_SRCCHAR(from_file);
current_input.state = INPUTSTATE_BOMKLUGE;
return 0xef;
}
} else {
// second byte does not match, so return first byte and remember to re-process second:
IF_WANTED_REPORT_SRCCHAR(0xef);
IF_WANTED_REPORT_SRCCHAR(from_file);
current_input.state = INPUTSTATE_AGAIN;
return 0xef;
}
BUG("InputBOM", 0);
}
IF_WANTED_REPORT_SRCCHAR(from_file);
//TODO - check for bogus/malformed BOM (0xef 0xbb 0xbf as UTF-8-encoded 0xfeff) and ignore?
// check for hashbang line and ignore
if (from_file == '#') {
// remember to skip remainder of line
@ -181,6 +209,10 @@ static char get_processed_from_file(void)
}
current_input.state = INPUTSTATE_AGAIN;
break;
case INPUTSTATE_BOMKLUGE:
// send second byte of non-BOM and then the byte read from file
current_input.state = INPUTSTATE_AGAIN;
return 0xbb;
case INPUTSTATE_NORMAL:
// fetch a fresh byte from the current source file
from_file = getc(current_input.u.fd);

2
src/version.h
View File

@ -9,7 +9,7 @@
#define RELEASE "0.97" // update before release FIXME
#define CODENAME "Zem" // update before release
#define CHANGE_DATE "22 Aug" // update before release FIXME
#define CHANGE_DATE "23 Aug" // update before release FIXME
#define CHANGE_YEAR "2024" // update before release
//#define HOME_PAGE "http://home.pages.de/~mac_bacon/smorbrod/acme/"
#define HOME_PAGE "http://sourceforge.net/p/acme-crossass/" // FIXME