1. exit(1) with an error message on stderr if an I/O error occurs.
1a. To work properly when built into /bin/sh sprinkle clearerr() at
appropriate places.
2. Verify that when a 'X data value is used with one of the numeric
conversions, that nothing follows the 'X'. It used to be unclear
in the standard whether this was required or not, it is clear that
with numeric conversions the entire data value must be used, or an
error must result. But with string conversions, that isn't the case
and unused parts are simply ignored. This one is a numeric conversion
with a string value, so which applies? The standard used to contain
an example of '+3 being converted, producing the same as '+ ignoring
the '3' with no mention of any error, so that's the approach we adopted,
The forthcoming version now explicitly states that an error would also
be generated from that case, as the '3' was not used by the numeric
conversion.
2a. We support those conversions with floating as well as integer conversions,
as the standard used to suggest that was required (but it makes no sense,
the values are always integers, printing them in a floating format is
dumb). The standard has been revised to make it clear that only the
integer numeric conversions %d %u %x (etc) are supposed to handle the 'X
form of data value. We still allow it with the floating formats as an
extension, for backward compat, just in case someone (other than the ATF
tests) is using it. It might go away.
2b. These formats are sypposed to convert 'X where 'X' is a character
(perhaps multibyte encoded) in the current LC_CTYPE locale category.
We don't handle that, only 1 byte characters are handled currently.
However the framework is now there to allow code to (one hopes, easily)
be added to handle multi-byte locales. (Note that for the purposes of
#2 above, 'X' must be a single character, not a single byte.)
worlds, as when the first arg (which should be the format) contains
no % conversions, and there are more args, the results are unspecified
(according to POSIX).
We can use this so the previous usage
printf -- format arg...
(which is stupid, and pointless, but used to work) continues to
simply ignore the -- (unspecified results mean we can do whatever
feels good...)
This brings back the #if 0'd block from the previous modification
(so there is no longer anything that needs cleaning up later) but runs
the getopt() loop it contained only when there are at least 2 args
(so any 1 arg printf always uses that arg as the format string,
whatever it contains, including just "--") and also only when the
first (format) arg contains no '%' characters (which guarantees no %
conversions without needing to actually parse the arg). This is the
(or a) "unspecified results" case from POSIX, so we are free to do
anything we like - including assuming that we might have options
(we don't) and pretending to process them.
(it doesn't - that is, shouldn't) which includes processing -- as an
"end of options". The first arg is (always) the format string.
Remove call to getopt() (but still do associated changes to argc/argv)
Note: for now this is #if 0's out instead of being deleted, the old
code should be fully removed sometime soon.
Problem pointed out on tech-userlevel by Thierry Laronde.
When printf is running builtin in a sh, global vars aren't reset to
0 between invocations. This affects "rval" which remembers state
from a previous %b \c and thereafter always exits after the first
format conversion, until we get a conversion that generates an
error (which resets the flag almost by accident)
printf %b abc\\c
abc (no \n)
printf %s%s hello world
hello (no \n, of course, no world ...)
printf %s%s hello world
hello
printf %s%s hello world
hello
printf %d hello
printf: hello: expected numeric value
0 (no \n)
printf %s%s hello world
helloworld (no \n, and we are back!)
This affects both /bin/sh and /bin/csh (and has for a very long time).
XXX pullup -8
POSIX requires that signed numbers (strings preceded by '+' or '-')
be allowed as inputs to all of the integer format conversions, including
those which treat the data as unsigned.
Hence we do not need a variant function whose only difference from its
companion is to reject strings starting with '-' - instead we use
the primary function (getintmax()) for everything and remove getuintmax().
Minor update to the man page to indicate that the arg to all of the
integer conversions (diouxX) must be an integer constant (with an
optional sign) and to make it blatantly clear that %o is octal and
%u is unsigned decimal (for some reason those weren't explicitly stated
unlike d i x and X). Delete "respectively", it is not needed (and does
not really apply).
XXX pullup -8
Avoid running off into oblivion when a format string,
or arg to a %b conversion ends in an unescaped backslash.
Patch from Leo slightly modified by me.
- ansification
- format of output of jobs command (etc)
- job identiers %+, %- etc
- $? and $(...)
- correct quoting of output of set, export -p and readonly -p
- differentiation between nornal and 'posix special' builtins
- correct behaviour (posix) for errors on builtins and special builtins
- builtin printf and kill
- set -o debug (if compiled with DEBUG)
- cd src obj (as ksh - too useful to do without)
- unset -e name, remove non-readonly variable from export list.
(so I could unset -e PS1 before running the test shell...)
andvar