Re: [Json] JSON Log file encoding JSON-L

[Nico Williams <[e-mail]>]

On Wed, May 7, 2014 at 4:18 PM, Phillip Hallam-Baker <[e-mail]> wrote:
> On Wed, May 7, 2014 at 5:04 PM, Nico Williams <[e-mail]> wrote:
>> On Wed, May 7, 2014 at 3:46 PM, Phillip Hallam-Baker <[e-mail]> wrote:
>>> If I open up a file with an RS character in it using notepad, how is
>>> it going to look?
>>
>> Notepad, Wordpad, my Unix shells, my Unix $EDITOR -- all display it as
>> a space or not at all.
>
> So how can I look at a file and know if it is a well formed sequence or not?

Use a JSON text processor, like jq(1).  (I'll have to update jq to
teach it to ignore RS between texts at the top-level.)

>>> Since RS is effectively an obsolete character the best I would expect
>>> is to see ? all over the place.
>>
>> Nope.  And even so, who cares?  It's a log file, you'll not be editing it.
>
> I open log files with an editor very frequently.

First, try it.  Second, what?  Why edit logfiles?  I've been dealing
with logfiles for decades and I've never had to edit one before.  What
are editing them for?

> It is easy enough to detect the sequence '}' *ws* LF *ws '{', its a
> four state FSR. Why complicate things with characters that will cause
> detours through untested code paths?

It's only necessary to know anything about RS when recovering from a
failure to parse a presumably-truncated entry in the log.  And even
then, we can simply seek forward to the next RS or NL and try parsing
whatever's there.

Nico
--