On Thu, Oct 11, 2012 at 12:29 PM, Jeroen Demeyer <[email protected]> wrote:
> On 2012-10-11 21:24, Volker Braun wrote:
>> As I wrote on the trac ticket before, there is virtually no difference
>> between "source foobar" in the shell and "import foobar" in Python, both
>> will happily pull in somebody else's code. And nobody would dream of
>> compiling a shared library into /tmp and then set LD_LIBRARY_PATH=/tmp
>> to load it (in fact this might happen behind the scenes if you were to
>> use libtool to compile something in /tmp). I don't see whats different
>> if you put a Python script in /tmp, thats just ill-advised.
>
> As I already said, I do think that Python is to blame here because doing
>
>   import some_module
>
> is an absolutely normal Python construct.

But putting Python files in /tmp and running them is not a normal
thing to do, any more than compiling shared libraries in /tmp. BTW,
it's not the Python testsuite that puts files here, it's us, which
should be fixed by #12415 (which is a much bigger issue).

> There is even no easy way to
> specify an absolute file path for "some_module".  In bash, people are
> not normally going to do
>
>   source some_file_not_under_my_control.sh
>
> (I would *always* specify a path for "source").

I've seen "source sibling_file" and "source
$path_of_this_file/sibling_file many times, how else are you going to
split functionality across multiple bash scripts? Such scripts
obviously shouldn't be executed from, e.g.,  /tmp.

- Robert

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
Visit this group at http://groups.google.com/group/sage-devel?hl=en.


Reply via email to