[test] Avoid unportable echo in Other/lit-quoting.txt

[ro]

Summary

LLVM :: Other/lit-quoting.txt currently FAILs on Solaris:

llvm/test/Other/lit-quoting.txt:8:9: error: CHECK2: expected string not found in input
CHECK2: {{^a\[b\\c$}}
        ^
<stdin>:1:1: note: scanning from here
a[b
^

This happens because echo with backslashes or special characters is unportable, as
extensively documented in the Autoconf manual. In the case at hand,
echo 'a[b\c' yields a[b\c on Linux, but a[b (no newline) on Solaris.

This patch fixes this by using the portable alternative suggested in the Autoconf manual.

Tested on amd64-pc-solaris2.11, sparcv9-sun-solaris2.11, and x86_64-pc-linux-gnu.

[mstorsjo]

I'm surprised to see that this actually seems to pass in the Windows CI environment (even if there are plenty of other unrelated failures in the CI run). On Windows, these tests aren't executed by a unix shell, but by the python lit internal shell, and echo is one of the commands handled internally there, but printf isn't.

[rnk]

... as extensively documented in the Autoconf manual ...

I hope I never have the joy of reading it. :)

In D108031#2944321, @mstorsjo wrote:

I'm surprised to see that this actually seems to pass in the Windows CI environment (even if there are plenty of other unrelated failures in the CI run). On Windows, these tests aren't executed by a unix shell, but by the python lit internal shell, and echo is one of the commands handled internally there, but printf isn't.

So, if I understand, the printf variant is actually a *better* test than the echo test because lit will actually call out to an external program. At least, I do believe this version does test the codepath that we care about, right? If you agree, I think this is fine.

[mstorsjo]

In D108031#2944367, @rnk wrote:

In D108031#2944321, @mstorsjo wrote:

I'm surprised to see that this actually seems to pass in the Windows CI environment (even if there are plenty of other unrelated failures in the CI run). On Windows, these tests aren't executed by a unix shell, but by the python lit internal shell, and echo is one of the commands handled internally there, but printf isn't.

So, if I understand, the printf variant is actually a *better* test than the echo test because lit will actually call out to an external program. At least, I do believe this version does test the codepath that we care about, right? If you agree, I think this is fine.

Hmm, I'm pretty sure I tested these to make sure they did test the thing that mattered (invoking external executables - and that the test did fail before applying the fix).

In any case, I checked now, and both GnuWin (which I think some buildbots use?) and Git Bash (which the CI runner uses) contain a printf.exe, and it does seem to work with GnuWin's printf.exe too, so I guess it should be fine then.

[mstorsjo]

I guess I shouldn't be blocking this, so as this seems to be working on Windows, LGTM.

llvm/test/Other/lit-quoting.txt

-RUN: echo "\"" | FileCheck %s --check-prefix=CHECK1
-RUN: echo '"' | FileCheck %s --check-prefix=CHECK1
-RUN: echo 'a[b\c' | FileCheck %s --check-prefix=CHECK2
-RUN: echo "a[b\\c" | FileCheck %s --check-prefix=CHECK2
-RUN: echo 'a\b\\c\\\\d' | FileCheck %s --check-prefix=CHECK3
-RUN: echo "a\\b\\\\c\\\\\\\\d" | FileCheck %s --check-prefix=CHECK3
+RUN: printf "%%s\n" "\"" | FileCheck %s --check-prefix=CHECK1
+RUN: printf "%%s\n" '"' | FileCheck %s --check-prefix=CHECK1
+RUN: printf "%%s\n" 'a[b\c' | FileCheck %s --check-prefix=CHECK2
+RUN: printf "%%s\n" "a[b\\c" | FileCheck %s --check-prefix=CHECK2
+RUN: printf "%%s\n" 'a\b\\c\\\\d' | FileCheck %s --check-prefix=CHECK3
+RUN: printf "%%s\n" "a\\b\\\\c\\\\\\\\d" | FileCheck %s --check-prefix=CHECK3
 CHECK1: {{^"$}}
 CHECK2: {{^a\[b\\c$}}
 CHECK3: {{^a\\b\\\\c\\\\\\\\d$}}
 
 On Windows, with MSYS based tools, the following commands fail though:
-RUNX: echo 'a[b\c\\d' | FileCheck %s --check-prefix=CHECK4
-RUNX: echo "a[b\\c\\\\d" | FileCheck %s --check-prefix=CHECK4
+RUNX: printf "%%s\n" 'a[b\c\\d' | FileCheck %s --check-prefix=CHECK4
+RUNX: printf "%%s\n" "a[b\\c\\\\d" | FileCheck %s --check-prefix=CHECK4
 CHECK4: {{^a\[b\\c\\\\d$}}