Re: zaml 0.071 bug

[Markus]

On Sun, 2008-12-14 at 01:43 -0800, Ian McIntosh wrote:
> A string like "- - (text) - -" seems to end up as an array:
>
> "title": - - (text) - -

Yep. The problem shows up in the new "proposed" branch too. I'm
inclined to fix it in proposed and call proposed the new main. But I
haven't heard anyone else's thoughts on the branch.

Thoughts?

-- Markus

[Jesse Hallett]

I think 'proposed' is ready to go.

Did you see my fixes for serializing "true" and "false"? That was a related bug.

On Dec 14, 2008 8:37 AM, "Markus" <mar...@reality.com> wrote:

On Sun, 2008-12-14 at 01:43 -0800, Ian McIntosh wrote: > A string like "- - (text) - -" seems to en...

[Markus]

> Did you see my fixes for serializing "true" and "false"? That was a
> related bug.
>

Yes. There's a minor design tension at work here. On the one hand,
there's the push for clean, human-friendly output. That suggests
omitting needless disambiguation syntax. But there's an even stronger
push towards correctness.

The easiest was to produce demonstrably correct results is to always
produce the disambiguation tinsel, but this clutters the output
considerably in conflict with the aforementioned goal. So we work at
charting out the boundary between the cases, hoping to produce all (and
only) the clutter we need.

The problem is (or eventually will be) that the process of determining
the absolute cleanest way to serialize every possible shape and flavor
of data may run is head-on into conflict with the third and forth goals:
clean code and fast execution.

I've identified a few other edge cases that may, I suspect, be
problematic. I'll try to work up a revised string output routine later
today & pass it on. I'm leaning towards a more conservative approach,
with correctness trumping cleanness.

--Markus