File: can-execute.txt

package info (click to toggle)
llvm-toolchain-11 1%3A11.0.1-2
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 995,808 kB
  • sloc: cpp: 4,767,656; ansic: 760,916; asm: 477,436; python: 170,940; objc: 69,804; lisp: 29,914; sh: 23,855; f90: 18,173; pascal: 7,551; perl: 7,471; ml: 5,603; awk: 3,489; makefile: 2,573; xml: 915; cs: 573; fortran: 503; javascript: 452
file content (23 lines) | stat: -rw-r--r-- 839 bytes parent folder | download | duplicates (16) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
 
REQUIRES: can-execute
REQUIRES: shell

This tests that we abstract two peculiarities of unix in can_execute:

* Directories are executable, but we don't want to try to execute them.
* For shell scripts, we also need to be able to read them.

The PATH is constructed such that 'not' will first find a directory named
TestProg, then a file with executable bit but not readable and finally a
shell script which always returns false, which is what it actually tries to
execute.

If we want, it is probably OK to change the semantics of can_execute and this
test, but for now this test serves as a reminder to audit all the callers if
we do that.

RUN: rm -rf %t && mkdir -p %t
RUN: cp -f %S/Inputs/TestProg/TestProg %t/TestProg
RUN: chmod 111 %t/TestProg
RUN: export PATH=%S/Inputs:%t:%S/Inputs/TestProg:$PATH
RUN: not TestProg
RUN: rm -rf %t