On Tue, 30 Aug 2022 14:09:40 +0200
Theo Buehler <t...@theobuehler.org> wrote:
> On Tue, Aug 30, 2022 at 01:01:47PM +0200, YASUOKA Masahiko wrote:
>> On Tue, 30 Aug 2022 11:56:53 +0200
>> Claudio Jeker <cje...@diehard.n-r-g.com> wrote:
>> > On Tue, Aug 30, 2022 at 11:18:01AM +0200, YASUOKA Masahiko wrote:
>> >> @@ -423,11 +423,13 @@ uu_decode(void)
>> >> }
>> >> }
>> >>
>> >> +#define ROUNDDOWN(x,y) (((x)/(y)) * (y))
>> >> +
>> >> static int
>> >> base64_decode(void)
>> >> {
>> >> int n;
>> >> - char inbuf[PATH_MAX];
>> >> + char inbuf[ROUNDDOWN(PATH_MAX, 4) + 1];
>> >> unsigned char outbuf[PATH_MAX * 4];
>> >>
>> >> for (;;) {
>> >>
>> >
>> > The fix is right but I wonder why is this code using PATH_MAX for a buffer
>> > size that has nothing to do with a file system path?
>>
>> It is a mystery of the history?
>>
>>
>> https://github.com/sergev/4.4BSD-Lite2/blob/master/usr/src/usr.bin/uudecode/uudecode.c#L92
>>
>> MAXPATHLEN was replaced to PATH_MAX in 2015.
>>
>> ok?
>>
>> Index: usr.bin/uudecode/uudecode.c
>> ===================================================================
>> RCS file: /cvs/src/usr.bin/uudecode/uudecode.c,v
>> retrieving revision 1.27
>> diff -u -p -r1.27 uudecode.c
>> --- usr.bin/uudecode/uudecode.c 28 Jun 2019 13:35:05 -0000 1.27
>> +++ usr.bin/uudecode/uudecode.c 30 Aug 2022 10:54:17 -0000
>> @@ -190,7 +190,7 @@ decode2(void)
>> void *handle;
>> struct passwd *pw;
>> struct stat st;
>> - char buf[PATH_MAX];
>> + char buf[BUFSIZ];
>>
>> base64 = 0;
>> /* search for header line */
>> @@ -342,7 +342,7 @@ uu_decode(void)
>> {
>> int i, ch;
>> char *p;
>> - char buf[PATH_MAX];
>> + char buf[BUFSIZ];
>>
>> /* for each input line */
>> for (;;) {
>> @@ -427,8 +427,8 @@ static int
>> base64_decode(void)
>> {
>> int n;
>> - char inbuf[PATH_MAX];
>> - unsigned char outbuf[PATH_MAX * 4];
>> + char inbuf[BUFSIZ];
>
> You need an extra byte for the terminating NUL to fix the problem you
> wanted to fix:
>
> char inbuf[BUFSIZ + 1];
>
> I'm not sure if we can rely on BUFSIZ to be a multiple of 4, I can't
> find anything that requires that. I think it's better to use a local
> #define with a comment as Claudio suggested.
Ah, I showed the diff separated from first one.
I'm sorry for confusing. The following diff is combined.
>
>
>> + unsigned char outbuf[BUFSIZ * 4];
>>
>> for (;;) {
>> switch (get_line(inbuf, sizeof(inbuf))) {
>>
Index: usr.bin/uudecode/uudecode.c
===================================================================
RCS file: /cvs/src/usr.bin/uudecode/uudecode.c,v
retrieving revision 1.27
diff -u -p -r1.27 uudecode.c
--- usr.bin/uudecode/uudecode.c 28 Jun 2019 13:35:05 -0000 1.27
+++ usr.bin/uudecode/uudecode.c 30 Aug 2022 12:15:56 -0000
@@ -190,7 +190,7 @@ decode2(void)
void *handle;
struct passwd *pw;
struct stat st;
- char buf[PATH_MAX];
+ char buf[BUFSIZ];
base64 = 0;
/* search for header line */
@@ -342,7 +342,7 @@ uu_decode(void)
{
int i, ch;
char *p;
- char buf[PATH_MAX];
+ char buf[BUFSIZ];
/* for each input line */
for (;;) {
@@ -423,12 +423,14 @@ uu_decode(void)
}
}
+#define ROUNDDOWN(x,y) (((x)/(y)) * (y))
+
static int
base64_decode(void)
{
int n;
- char inbuf[PATH_MAX];
- unsigned char outbuf[PATH_MAX * 4];
+ char inbuf[ROUNDDOWN(BUFSIZ, 4) + 1];
+ unsigned char outbuf[BUFSIZ * 4];
for (;;) {
switch (get_line(inbuf, sizeof(inbuf))) {
