Re: [Qemu-block] [PATCH v4] tests: Avoid non-portable 'echo -ARG'

[Kevin Wolf]

Am 08.08.2017 um 16:29 hat Eric Blake geschrieben:
> On 08/08/2017 08:54 AM, Kevin Wolf wrote:
> > Am 03.07.2017 um 20:09 hat Eric Blake geschrieben:
> >> POSIX says that backslashes in the arguments to 'echo', as well as
> >> any use of 'echo -n' and 'echo -e', are non-portable; it recommends
> >> people should favor 'printf' instead.  This is definitely true where
> >> we do not control which shell is running (such as in makefile snippets
> >> or in documentation examples).  But even for scripts where we
> >> require bash (and therefore, where echo does what we want by default),
> >> it is still possible to use 'shopt -s xpg_echo' to change bash's
> >> behavior of echo.  And setting a good example never hurts when we are
> >> not sure if a snippet will be copied from a bash-only script to a
> >> general shell script (although I don't change the use of non-portable
> >> \e for ESC when we know the running shell is bash).
> >>
> 
> >> +++ b/tests/multiboot/run_test.sh
> >> @@ -26,7 +26,7 @@ run_qemu() {
> >>      local kernel=$1
> >>      shift
> >>
> >> -    echo -e "\n\n=== Running test case: $kernel $@ ===\n" >> test.log
> >> +    printf %b "\n\n=== Running test case: $kernel $@ ===\n\n" >> test.log
> >>
> >>      $QEMU \
> >>          -kernel $kernel \
> > 
> > Not completely sure why, but this broke the test with whitespace changes
> > like this:
> > 
> > -=== Running test case: mmap.elf -m 1.1M ===
> > +=== Running test case: mmap.elf -m1.1M ===
> 
> I guess that means I'm not regularly running tests/multiboot?  Is it not
> part of 'make check' or qemu-iotests?

The problem is that it needs an i386 compiler to build the test kernels
(and qemu-system-i386 or qemu-system-x86_64 binaries to execute them).

I guess we could check these conditions, though, and skip the test if we
can't produce i386 binaries.

> Ah, I see the problem, and it's insidious.  We're using "address@hidden", but
> want to be using "...$*...".  $@ causes multiple arguments to be passed,
> but printf %b is not concatenating those arguments; while $* uses only a
> single argument.  We didn't notice it with echo -e, because echo inserts
> a space between multiple arguments, just as you'd get a space with $*.

The thing that completely confused me here is that printf doesn't just
ignore additional arguments as I would have expected, but just starts
over with the format string, so that it does kind of work with multiple
arguments and fails only subtly.

Kevin

signature.asc
Description: PGP signature