File: can-execute.txt

package info (click to toggle)
llvm-toolchain-9 1%3A9.0.1-16.1
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 882,388 kB
  • sloc: cpp: 4,167,636; ansic: 714,256; asm: 457,610; python: 155,927; objc: 65,094; sh: 42,856; lisp: 26,908; perl: 7,786; pascal: 7,722; makefile: 6,881; ml: 5,581; awk: 3,648; cs: 2,027; xml: 888; javascript: 381; ruby: 156
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