[darcs-users] darcs patch: tests/*: refactor remaining IFS calls to lib

[Eric Kow]

On Mon, Dec 01, 2008 at 11:56:47 -0500, Gwern Branwen wrote:
> 1) Only lib callers are affected.
> 2) Should any of our scripts be calling read?

Hmm... is that the only way for them to be affected by the IFS variable?
Clearly it does *something* as we have to set it on some occasion or
another, seemingly to control what happens when we use a backtick.

> 3) Each test should be running in its own shell or subshell for
> maximum separation.

This is currently the case.  My worry was not about leakage from one
test to the other, but through all tests using the same library.

> 4) We can conditionally set IFS on Windows, ie.
> if echo $OS | grep -i windows; then
>   IFS=''
> fi

I think I would be a bit more comfortable with that.

Well, anyway, I guess I'll apply this patch if you still think this is a
good idea, maybe with the above code to limit the effects to Windows.

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : http://lists.osuosl.org/pipermail/darcs-users/attachments/20081202/d34b1a8e/attachment.pgp