- The input loop control that I changed yesterday to:
while (mbslength >= 0) {
There are circumstances where this causes an extra \000 to
be added at the end of some tests. This error was showing
in my own tests here, but I did not notice it yesterday.
(I really need to add my tests to the test suite, catching
every error by eye is hard.) To fix, I've now changed the
code to increment mbslength only if mbslength == 1 to start
with. (Note that this check for "== 1" is why the arg to
strvisx() in vis(1) must be 1, not mbilen.)
- The cast sequence when manually inserting bytes after a
multibyte conversion error:
*src = (wint_t)(u_char)*mbsrc;
is wrong. This is causing problems in the case when an
8859-1 input string is processed in the UTF-8 locale.
It needs to be:
*src = (wint_t)*mbsrc;
Without the (u_char) all the locale mismatch combinations
then work.
- The code:
if (mblength < len)
len = mblength;
needs to be there. It resets len for the single character
input case after we've actually processed two input
characters (c and nextc) because we incremented mbslength
at the start of the loop. Without this code, single
character conversions end up with a \000 or other byte
appended.
the soft float or hard float ABI is being used. However, if there isn't a
FPU that can't be done. So only save/restore them if a FPU is present. When
libc initializes, it does a sysctl to determine if there is a FPU and stores
the result which _setjmp/setjmp uses. If there was a FPU, the magic in the
jmp_buf is changed to reflect that the VFP registers were saved. longjmp uses
the magic to determine if it needs to restore the VFP registers.
the *at system calls on November 18th of last year. Reasons to revert
it include:
- it is incorrect in a whole variety of ways (but fortunately, one
of them is that the missing and improper permission checks have
no net effect);
- it was committed without review or discussion;
- core ruled that all the new O_* flags pertaining to the *at calls
needed to wait until their semantics could be clarified.
manu was asked to revert it on these grounds but has ignored the request.
I have left O_SEARCH defined and visible and made open() explicitly
ignore it. This way, most code that tries to use it will continue to
build and run. I've also arranged lib/libc/c063/t_o_search.c so that
the tests that make use of the O_SEARCH semantics will disappear until
O_SEARCH comes back, and fixed some mistakes and/or incorrect hacks
that were causing some of these to succeed despite the broken O_SEARCH
implementation.
some wording related to the *at form... but it needed a general overhaul.
Add some missing errors for the *at form... plus EINVAL for the
traditional form for when you pass a bogus check mode.
Note that the AT_EACCESS flag is useless and strengthen the security
warning.
avoid saying "we", fix typos, etc.
XXX: groff and mandoc handle the URLs in here differently but both
XXX: mangle them, apparently because they're wider than one line.
XXX: Does anyone know how to fix this or do we need an URL shortener
XXX: on www.NetBSD.org?
