On Sat, May 25, 2013 at 05:59:37PM +0200, Moritz Heidkamp wrote:
> Mario Domenech Goulart writes:
> > Maybe "readlink -f" ?
>
> OK, Peter found that even that is not portable so I changed the test
> itself to do the path canonicalization. Turns out that while the posix
> unit's read-symbolic-link
Mario Domenech Goulart writes:
> Maybe "readlink -f" ?
OK, Peter found that even that is not portable so I changed the test
itself to do the path canonicalization. Turns out that while the posix
unit's read-symbolic-link has a CANONICALIZE option it doesn't quite
behave like the readlink(1) progr
Mario Domenech Goulart writes:
>> Ah good catch, but unfortunately OpenBSD does not expose realpath
>> as a shell command (or binary utility). Is there a way to do this
>> portably? John, do you know?
>
> Maybe "readlink -f" ?
Thanks, that works, as well! Attached is an updated patch.
Moritz
>Fr
On Sat, 25 May 2013 02:01:24 +0200 Christian Kellermann
wrote:
> * Moritz Heidkamp [130525 01:49]:
>> The private repository path tests didn't work when run from inside a
>> path containing symlinks because runtests.sh didn't expand symlinks
>> while the -private-repository mechanism does. This
* Moritz Heidkamp [130525 01:49]:
> The private repository path tests didn't work when run from inside a
> path containing symlinks because runtests.sh didn't expand symlinks
> while the -private-repository mechanism does. This lead the test
> assertion which compares the two paths to fail.
Ah go
The private repository path tests didn't work when run from inside a
path containing symlinks because runtests.sh didn't expand symlinks
while the -private-repository mechanism does. This lead the test
assertion which compares the two paths to fail.
>From 464869bde8c7d70af42cc7b9530181bb78698370