On Tue, 26 Sep 2023 12:41:47 +0900
Takashi Yano wrote:
> Hi,
> 
> I noticed that the following test case crashes at printf() with current
> newlib.
> 
> 
> #include <stdio.h>
> #include <wchar.h>
> #include <locale.h>
> 
> int main()
> {
> 	setlocale(LC_ALL, "C.UTF-8");
> 	wprintf(L"%ls\n", L"aaaa"); /* or fwide(stdout, 1); */
> 	printf("%ls\n", L"bbbb");  /* <--- crash here */
> 	return 0;
> }
> 
> 
> I looked into this problem and found the cause.
> 
> A narrow char string which can be odd bytes in length is cast into
> a wide char string which should be even bytes in length in __sprint_r/
> __sfputs_r based on the __SWID flag. As a result, if the length is
> odd bytes, the reading buffer runs over the buffer length, which causes
> a crash. If the length is even bytes, crash does not happen, but garbage
> is printed. This hapens if printf("%ls\r\n", L"bbbb"); is used instead.
>                                       ^^
> 
> The same issue seemed to be reported ten years ago.
> https://sourceware.org/pipermail/newlib/2013/010831.html
> 
> I have built a patch attached for this issue.
> 
> With this patch, __sfputs_r/__sprint_r is split into two versions, one
> is for vfprintf which does not handle wide string, and the other (newly
> introduced __sfputws_r/__swprin_r) is for vfwprintf which handles wide
> string. Please note that fprintf gets working for wide orient stream
> just like BSD libc, which behaves differently from GNU libc.
> 
> This patch also fixes nano-vfprintf.c as well as vfprintf.c/vfwprintf.c
> in the same manner.

v2: Remove __sprint_r from nano-vfprintf.c which does not seem to be used
    anymore.

-- 
Takashi Yano <takashi.yano@nifty.ne.jp>