Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 15 Sep 2016 12:37:57 -0400
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: [PATCH] fix printf regression with alt-form octal,
 default precision

On Thu, Aug 04, 2016 at 04:07:40AM +0300, Dmitry V. Levin wrote:
> commit v1.1.5-10-gb91cdbe2bc8b626aa04dc6e3e84345accf34e4b1 that fixed
> behavior of printf with alt-form octal, zero precision, zero value,
> at the same time broke alt-form octal with default precision,
> e. g. printf("%#09o", 1).
> ---
>  src/stdio/vfprintf.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/src/stdio/vfprintf.c b/src/stdio/vfprintf.c
> index 2ecf769..ac2891c 100644
> --- a/src/stdio/vfprintf.c
> +++ b/src/stdio/vfprintf.c
> @@ -570,7 +570,14 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
>  			if (0) {
>  		case 'o':
>  			a = fmt_o(arg.i, z);
> -			if ((fl&ALT_FORM) && p<z-a+1) p=z-a+1;
> +			if (fl&ALT_FORM) {
> +				if (p >= 0 && p<z-a+1) {
> +					p=z-a+1;
> +				} else if (arg.i) {
> +					prefix+=5;
> +					pl=1;
> +				}
> +			}
>  			} if (0) {
>  		case 'd': case 'i':
>  			pl=1;

I think the attached simpler patch should work just as well. Adjusting
'p' to implement '#' was probably a bad idea since it makes the later
code paths think an explicit precision was specified when it was not.
"Add a prefix of '0' when p is too short to have a leading '0'
already" seems like safer logic.

I'm also attaching some test cases I ran.

Rich



View attachment "octal_alt_form.diff" of type "text/plain" (435 bytes)

View attachment "alt_form_octal.c" of type "text/plain" (404 bytes)

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.