[PATCH] Failed test: scm_rights-fd.test for path containing '>', '}' or non-printable characters

Jay Joshi jay.r.joshi100 at gmail.com
Wed Mar 23 14:14:24 UTC 2016


On Wed, Mar 23, 2016 at 1:24 AM, Dmitry V. Levin <ldv at altlinux.org> wrote:
> On Tue, Mar 22, 2016 at 09:52:55PM +0530, Jay Joshi wrote:
>> On Tue, Mar 22, 2016 at 5:29 AM, Dmitry V. Levin <ldv at altlinux.org> wrote:
>> > On Mon, Mar 21, 2016 at 05:18:10PM +0530, Jay Joshi wrote:
>> >> >From d718d0c8e94cab8a232caf13e282fdea29b9a8eb Mon Sep 17 00:00:00 2001
>> >> From: JayRJoshi <jay.r.joshi100 at gmail.com>
>> >> Date: Mon, 21 Mar 2016 17:15:11 +0530
>> >> Subject: [PATCH] fix pathname containing >,{ and other non-printables
>> >
>> > Good catch.
>> >
>> >> tests/scm_rights-fd.test: changed regexp for pathname in $rights
>> >> ---
>> >>  tests/scm_rights-fd.test | 2 +-
>> >>  1 file changed, 1 insertion(+), 1 deletion(-)
>> >>
>> >> diff --git a/tests/scm_rights-fd.test b/tests/scm_rights-fd.test
>> >> index 48c5028..469348b 100755
>> >> --- a/tests/scm_rights-fd.test
>> >> +++ b/tests/scm_rights-fd.test
>> >> @@ -54,7 +54,7 @@ run_strace -y -x -enetwork $args "$file"
>> >>  sample='\\xf1\\xf2\\xf3\\xf4\\xf5\\xf6\\xf7\\xf8\\xf9\\xfa\\xfb\\xfc\\xfd\\xfe\\xff'
>> >>  n='[1-9][0-9]*'
>> >>  msg='\{msg_name\(0\)=NULL, msg_iov\(1\)=\[\{"'"$sample"'", 15\}\],
>> >> msg_controllen='"$n"
>> >> -rights='\{cmsg_len='"$n"', cmsg_level=SOL_SOCKET,
>> >> cmsg_type=SCM_RIGHTS, \[4</dev/null>, 5</dev/zero>,
>> >> 6</[^}>]*/(A\\n){127}Z>\]\}'
>> >> +rights='(\{cmsg_len='"$n"', cmsg_level=SOL_SOCKET,
>> >> cmsg_type=SCM_RIGHTS, \[4</dev/null>, 5</dev/zero>,
>> >> 6</[^\0]*/(A\\n){127}Z>\]\}|\{cmsg_len='"$n"', cmsg_level=SOL_SOCKET,
>> >> cmsg_type=SCM_RIGHTS, \[4</dev/null>, 5</dev/zero>,
>> >> 6<(\\x[0-9a-f][0-9a-f])*\\x2f(\\x41\\x0a){127}\\x5a>\]\})'
>> >
>> > You don't have to clone the whole regexp if the only thing that has to be
>> > changed is the part of regexp related to 6<...>.
>> >
>> > BTW, I don't see how your change fixes the issue.
>>
>> First part of "or" was to match for path-name containing printable
>> characters only. For second part, string_quote while "x" flag set
>> print whole string in hexadecimal if at least one of the character is
>> is non-printable.
>
> So there are two bugs in the test, one is that '>' and '}' in $PWD are not
> handled properly, another one is that similar issue arises with those
> symbols in $PWD for which strace -x turns into hex mode.
>

Yes. when path-name contains non-printable, at that time usehex is
set, which turns hex mode on.

>> Updated patch is pasted. Am I missing something?
> [...]
>> -rights='\{cmsg_len='"$n"', cmsg_level=SOL_SOCKET,
>> cmsg_type=SCM_RIGHTS, \[4</dev/null>, 5</dev/zero>,
>> 6</[^}>]*/(A\\n){127}Z>\]\}'
>> +rights='\{cmsg_len='"$n"', cmsg_level=SOL_SOCKET,
>> cmsg_type=SCM_RIGHTS, \[4</dev/null>, 5</dev/zero>,
>> 6<(/[^\0]*/(A\\n){127}Z|\\x2f(\\x[0-9a-f][0-9a-f])*\\x2f(\\x41\\x0a){127}\\x5a)>\]\}'
>
> What did you mean by [^\0]* ?  It doesn't work for me, while simple .* does.
>

Yes, it doesn't make much sense. More appropriate expression would be
[^\0]+, so that at least path wouldn't start with "//". Although,
path-name containing "//" is valid. So, .* is better.




More information about the Strace-devel mailing list