[Fuzzer] make big-file-copy.test work with the internal shell #168658
Conversation
This patch uses several shell features not supported by the internal shell, such as $? to get the exit code of a command, and exit. This patch adjusts the test to work with the internal shell by using bash to run the actual command with a zero exit code to ensure the file is deleted, and python to propagate the exit code up to lit.
|
@llvm/pr-subscribers-compiler-rt-sanitizer Author: Aiden Grossman (boomanaiden154) ChangesThis patch uses several shell features not supported by the internal shell, such as $? to get the exit code of a command, and exit. This patch adjusts the test to work with the internal shell by using bash to run the actual command with a zero exit code to ensure the file is deleted, and python to propagate the exit code up to lit. Full diff: https://github.com/llvm/llvm-project/pull/168658.diff 1 Files Affected:
diff --git a/compiler-rt/test/fuzzer/big-file-copy.test b/compiler-rt/test/fuzzer/big-file-copy.test
index 9565ed4e0bca2..3008933794ae1 100644
--- a/compiler-rt/test/fuzzer/big-file-copy.test
+++ b/compiler-rt/test/fuzzer/big-file-copy.test
@@ -1,6 +1,6 @@
REQUIRES: darwin
UNSUPPORTED: ios
RUN: %cpp_compiler %S/BigFileCopy.cpp -o %t
-RUN: %run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; result=$?
+RUN: bash -c "%run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; echo -n $? > %t.result"
RUN: %run rm -f big-file.txt big-file-out.txt
-RUN: (exit $result)
+RUN: %python -c "import sys; sys.exit(%{readfile:%t.result})"
|
🐧 Linux x64 Test Results
|
| UNSUPPORTED: ios | ||
| RUN: %cpp_compiler %S/BigFileCopy.cpp -o %t | ||
| RUN: %run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; result=$? | ||
| RUN: bash -c "%run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; echo -n $? > %t.result" |
There was a problem hiding this comment.
This will no longer work with an external shell (same reason as the other test: outer shell expands $? before inner shell has executed anything), please make the internal shell the default at the same time you merge this.
Quite unexpectedly with an external shell this writes an empty file and thus below will execute sys.exit() (without passing a return code) which will always pass.
There was a problem hiding this comment.
Switched to using single quotes so this also works with the external shell.
| RUN: bash -c "%run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; echo -n $? > %t.result" | ||
| RUN: %run rm -f big-file.txt big-file-out.txt | ||
| RUN: (exit $result) | ||
| RUN: %python -c "import sys; sys.exit(%{readfile:%t.result})" |
There was a problem hiding this comment.
This is sort of ridiculous, no? If we are going out to bash anyway, could we do something like:
RUN: bash -c "%run %t -runs=1 -rss_limit_mb=4096 2>big-file-out.txt; result=$?; \
RUN: %run rm -f big-file.txt big-file-out.txt && exit $result"
There was a problem hiding this comment.
Good point. Updated to use something more similar to what you proposed to avoid all the indirection through files that is mostly pointless.
3 participants
This patch uses several shell features not supported by the internal shell, such as $? to get the exit code of a command, and exit. This patch adjusts the test to work with the internal shell by using bash to run the actual command with a zero exit code to ensure the file is deleted, and python to propagate the exit code up to lit.