On Wed, Apr 27, 2016 at 3:06 PM, Alex Shinn <xxxxxx@gmail.com> wrote:

So leaving this unspecified is probably the right thing, but I think
including another feature to say whether POSIX semantics were
used would have been useful. In the meantime it's simple enough
to test programmatically with your example - all of the backtracking
algorithms will get this wrong:

Would you like to include some mention of this issue in a revised version of the document? While we shouldn't change the meaning of the SRFI, it's okay to include a clarification or an obvious error correction. We could add a paragraph, and I would add something like "Revised to fix errata: 2016/4/27" to the Status section. (I could add the missing <h1> for the Status section in the process.)