Before checkin aeb0e0e9f1 we would only
invoke ofmt->symdef() for a common variable during pass 2 if that
variable had "special" in the declaration. That checkin
unintentionally changed that behavior.
That doesn't mean that the pre-existing behavior is right, but this
should at least fix the resulting regressions. This really warrants
more investigation.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
A so-called "generalized CRC" is a form of hash function based on a
table similar to the conventional bytewise software implementation of
CRC. For each byte in each data set, it contains a random
byte permutation table.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Small utility program to generate CRC tables in both C and Perl
formats. Note: we don't actually need a true CRC, which is what this
generates. We may want to consider going to a "generalized CRC",
meaning a randomly generated set of 8 bytewise permutations. These
have the advantage of not being linear, while having the same software
implementation as a true CRC.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
It isn't just invalid UTF-8 sequences that are affected; it is in fact
any high-bit bytes (we don't ever try to construct UTF-8 characters,
just deconstruct them.)
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
When computing the length of an octal escape, we need to do an
unsigned compare, otherwise we only allocate space for one character
for bytes in the \200..\377 range, which is obviously incorrect.
Reported-by: Ed Beroset <beroset@mindspring.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
There is no point in using sprintf(), and it adds the possibility of
either bugs due to the output not matching what the byte count loop is
expecting, or just cause people to freak out due to the notion that
"sprinf is unsafe".
Reported-by: Ed Beroset <beroset@mindspring.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
When computing the length of an octal escape, we need to do an
unsigned compare, otherwise we only allocate space for one character
for bytes in the \200..\377 range, which is obviously incorrect.
Reported-by: Ed Beroset <beroset@mindspring.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Strings returned by nasm_unquote() can contain NUL characters, which
will not be legal if then used as a C string. Create a general
function which looks for NUL characters in the string and issues an
error if they are found.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
There is no point in using sprintf(), and it adds the possibility of
either bugs due to the output not matching what the byte count loop is
expecting, or just cause people to freak out due to the notion that
"sprinf is unsafe".
Reported-by: Ed Beroset <beroset@mindspring.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Split out the global directives into a separate section, that allows
the switch() in the main code to be slightly faster.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Hash all directives, even the ones that are backend-specific,
and instead pass the backend an already-parsed directive number.
Furthermore, unify null functions across various backends.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
We fopen() the output file in common code but fclose() it in the
backend. This is bad for a variety of reasons:
1. it is generally an awkward interface to change ownership.
2. we should use ferror() to test for write errors, and that is
better done in common code.
3. it requires more code.
4. we still need to fclose() in common code during error handing.
Thus, move the fclose() of the output out of the backends, and add
fflush() so we can test ferror() on output.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
