WARNS=6

branches: 1.11.2; 1.11.8;
document non-literal format strings

Enable WARNS=4 by default for usr.bin, except for:
awk bdes checknr compile_et error gss hxtool kgetcred kinit
klist ldd less lex locale login m4 man menuc mk_cmds
mklocale msgc openssl rpcgen rpcinfo sdiff spell ssh
string2key telnet tn3270 verify_krb5_conf xlint

branches: 1.9.36;
WARNS=3, and remove comment about WFORMAT issues.

WARNS=2

More format string cleanup by sommerfeld.

branches: 1.6.2; 1.6.10;
remove WARNS?=1, cleanup .Nm usage

branches: 1.5.2;
"merge" lite-2. our printf is already kinda different...minor changes only.

Fix compiler warnings.

RCS ID police

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

Random minor wording and markup cleanups.

printf(1): misc tweaks

Move the advise about getopt's "--" to where the format is disussed.
Use .Vt for types.

printf(1): move \e description to its lexicographic place

Add %C format conversion and -L option to printf(1)

%C does what everyone always thought %c should do, but doesn't,
and operates rather like the %c conversion in printf(3) (to be
more precise, like %lc). It takes a code point integer value
in the current locale's LC_CTYPE and prints the character designated.

-L (this printf's first, and only, option) makes the floating conversions
use long double instead of double.

In the manual (printf.1) document both of those, and also be more
precise as to when things are affecting bytes, and when they're
manipulating characters (which makes no difference if LC_ALL=C).

branches: 1.37.2;
s/chracter/character/ in documentation.

branches: 1.36.4;

Don't describe "%%: in the format as a format conversion, it isn't,
it is an escaped literal '%' character. None of what applies to
format conversions applies to this.

branches: 1.35.2;
Add quotes around command substitution in the example
so that it works regardless of IFS and buts.
Requested by kre@

Drop redundant quoting in the nested printf example.

Use ASCII single quote.

Try to improve markup to get a nicer PostScript output.

While here edit slightly for consistency: get the octal character
ranges in order, use \num to refer to octal everywhere, etc.

PR standards/53563

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

Add the new formats to the list of format cracters (oops...)

Add support for F a and A formats (which go with the eEfgG formats
already supported.)

Correct a typo (off by one (key)) ...
There is no 'w' format, but there is an 'e'

branches: 1.27.2;
Remove Tn.

Explicitly mention use of -- for format strings beginning with -,
per PR 21970.

branches: 1.25.16; 1.25.22;
Change some standard exit status text to ".Ex -std"
From Eitan Adler in PR bin/47790.

branches: 1.24.6; 1.24.12; 1.24.22;
\\*(Pm -> \*(Pm

Use semantic markup

branches: 1.22.6;
Fix description of %b format. PR 39111 from Paul Goyette.

branches: 1.21.2;
Add a note about the stupid behaviour of "%c" format.

branches: 1.20.32;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

Use .Aq instead of \*[Lt]...\*[Gt]; sort SEE ALSO; add some more commas;
new sentence, new line; fix some mdoc nits.

Add missing comma.

Fixes from David Laight:
- 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...)

New sentence, new line.

Generate <>& symbolically. I'm avoiding .../dist/... directories for now.

Punctuation nits.

Whitespace cleanup.

Xr printf(9)

For commands and utilities, use EXIT STATUS rather than RETURN VALUES or
DIAGNOSTICS as appropriate (and documented in mdoc(7)).

\e generates an escape character but is not listed in the manpage. Fixed.

remove WARNS?=1, cleanup .Nm usage

"merge" lite-2. our printf is already kinda different...minor changes only.

alternate -> alternative, per PR 2643

RCS ID police

Changed floating point output format to be exactly the same (use same
text and macros) as those in the printf(3) manpage.
This fixes PR #1194, albeit differently.

The descriptions of the o and u formatting options were reversed (PR #785).

Changes required to make printf utility POSIX.2 compliant:
* Escape characters in the string needed to be processed as they were
encountered, otherwise a "\000" octal constant would prematurely
terminate the formatting string.
* Implemented the %b, SysV echo(1) compatibility, formatting directive.

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

Improve detection and diagnosis of invalid values for conversions.
(In particular, integer conversions contain no spaces, and must always
contain at least 1 digit, '' is not valid).

Correctly handle extracting wide chars from empty strings.

Fix a (probably would have rarely been seen) bug I installed yesterday.

It turns out that mbtowc() needs to include the terminating \0 in the
length arg passed to it, or it errors (EILSEQ) on a zero length (instead
of doing the sane thing and treating that the same as "\0" (treated as
being length 1). So, increase the length passed to mbtowc() by 1.
That makes no difference in the typical case, it is an upper limit on
the number of bytes to examine, and mbtowc() stops after it has
converted 1 character, so in the non "" input cases, nothing that
matters changes.

The rest of this you can skip if you like, not directly related to
this change...

Note: it is not clear to me what is correct here, POSIX looks to be
ambiguous, or strange anyway; in the RETURN VALUE section it says:

If s is not a null pointer, mbtowc() shall either return 0 (if s points
to the null byte), or return the number of bytes [...]

Further for the error possibilities it says:

[EILSEQ] An invalid character sequence is detected. In the POSIX locale
an [EILSEQ] error cannot occur since all byte values are valid
characters.

On the other hand our mbtowc(3) says:

There are special cases:

n == 0 In this case, the first n bytes of the array pointed to by
s never form a complete character. Thus, the mbtowc()
always fails.

Since EILSEQ is the only defined error for mbtowc() in POSIX, and
cannot happen (according to it) in the POSIX locale, that "always fails"
in our manual page looks dubious.

What actually happens in our mbtowc() in the POSIX locale, is that if
passed n==0 (and *s == '\0') mbtowc() returns 0 (that's good) but
also sets errno to EILSEQ (not so good - though this is not one of
the functions guaranteed to not alter errno if it doesn't fail).

In other locales it returns -1 (with errno == EILSEQ) when n == 0.
(Well, in some other locales anyway, I didn't go and test all of them).

Where POSIX gets weird, is that earlier it says:

At most n bytes of the array pointed to by s shall be examined.

If n == 0, then no bytes can be examined. In that case mbtowc()
cannot test whether s points to the null byte, even in the POSIX locale.

So it is unclear (to me) what should be returned in that case.

Add %C format conversion and -L option to printf(1)

%C does what everyone always thought %c should do, but doesn't,
and operates rather like the %c conversion in printf(3) (to be
more precise, like %lc). It takes a code point integer value
in the current locale's LC_CTYPE and prints the character designated.

-L (this printf's first, and only, option) makes the floating conversions
use long double instead of double.

In the manual (printf.1) document both of those, and also be more
precise as to when things are affecting bytes, and when they're
manipulating characters (which makes no difference if LC_ALL=C).

PR bin/58534 -- printf(1) source code comment fix

Update the comment near the start of main() in printf.c so it
explains what is really happening and why, rather than being a
whole bunch of incorrect BS about what posix does or doesn't require.

This changes comments only, NFC (should be no binary change at all).

Fix typo in comment.

Reported by Emanuele Torre in PR 58439.

branches: 1.54.6;
fix typo

Changes for POSIX conformance.

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.)

branches: 1.52.2;
make value an int to avoid all the casts and conversion warnings.

Change octal and hex parsing to not use strtoul so that they don't handle
'-'. From Martijn van Duren.
Also add a warning if the conversion fails (like the gnu printf does)

Amend the previous change: we can have (almost) the best of both
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.

Stop assuming that printf handles options in any way at all
(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.

Revert previous, it was based upon a misreading of the POSIX
spec. POSIX requires "as if by calling strtod()" which we
did already ... by calling strtod(). Go back to doing that.

Always convert input numbers (from the command line) in the C
locale, not as set in the environment. Conforms with POSIX spec.

A truly ancient bug found by Edgar Fuss

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

Printf's that support \e for escape all seem to also support \E.
Except us. Now we do as well.

Tighten syntax a little (no more %*4.*2d nonsense).
Include the format collected so far in "missing format char" err message.
Minor KNF and whitespace.

PR standards/53563

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

NFC: More KNF (remove () around returned constants).

NFC: whitespace & KNF.

Add support for F a and A formats (which go with the eEfgG formats
already supported.)

Avoid printing error messages twice when an invalid
escape sequence (\ sequence) is present in an arg to a %b
conversion.

From leot@ on tech-userlevel:

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.

branches: 1.37.8; 1.37.14; 1.37.16;
fix some error handling.

branches: 1.36.6; 1.36.8; 1.36.12;
WARNS=6

branches: 1.35.4; 1.35.10;
support grouping format.

Avoid segv on "printf '%*********s' 666", from Maksymilian Arciemowicz

branches: 1.33.4; 1.33.8; 1.33.10;
Remove the \n and tabs from the __COPYRIGHT() strings.
Tweak to use a consistent format.

branches: 1.32.4;
detect more errors from printf/malloc.

Remember to consume input bytes when processing '\0nnn' for %b formats

branches: 1.30.2;
- KNF, WARNS=3, pass lint.
- Simplify octal parsing code.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

Revert previous. 'None' means that the "Utility Syntax Guidlines" apply.

Remove getopt() loop, IEEE 1003.1 doesn't say that printf(1) should conform
to the "Utility Syntax Guidlines".
Fixes PR 21970.

Fix the output of NUL bytes within %b formats.
(Approved by Christos)

Fixes from David Laight:
- 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...)

Complete declaration of progprintf() to fix build problem in csh(1).

Remove #ifdef __STDC__. De-__P() and ANSIfy. Fix a prototype mismatch
uncovered by this.

Change to use {u,}intmax_t internally (was: (unsigned) long).

brace pollution, and char -> unsigned char

include unistd

add <unistd.h> to fix compiler warning

s/index/strchr

"merge" lite-2. our printf is already kinda different...minor changes only.

Fix compiler warnings.

lint and KNF changes. (mostly casting returns to void to quiet lint.)

RCS ID police

Simplify conversion of "quoted" numeric arguments.

Code to check if conversion (by strtol(), strtoul(), or strtod()) was
identical, so I moved it into its own function.

Add and use getulong() to handle %u, %o, %x & %X formatting directives.
It was using getlong(), which caused values larger than LONG_MAX to be
truncated to LONG_MAX.
As recommended by 1003.2, print warning messages when argument cannot be
converted to value or is out of range.

Handle format strings error correctly.

Error in hextobin() macro messed up hex escape constants.

Oops! get rid of the free(), mklong()'s buffer no longer malloc()'d.

Return from main() if a \c escape is encountered in a %b string (was an exit()).
Use macro constants for "skip1" and "skip2" instead of assigning them
each loop iteration.
Reformat the multi-case entries in the "big switch" so the lines don't wrap.

Move all the code from do_printf() into do-while loop in main(). I need
to be able to return from main() when a "\c" in a %b string is encountered.

Merged in most of the changes from 4.4 necessary to make printf a sh
and csh builtin --- still need to handle the one remaining exit() in
the SysV escape string handling code.

Changes required to make printf utility POSIX.2 compliant:
* Escape characters in the string needed to be processed as they were
encountered, otherwise a "\000" octal constant would prematurely
terminate the formatting string.
* Implemented the %b, SysV echo(1) compatibility, formatting directive.

Add RCS identifiers.

Cleanup for GCC 2.

branches: 1.1.1;
Initial revision