On Thu, Apr 27, 2017 at 9:54 PM, Shane Curcuru <[email protected]> wrote:
> Sam Ruby wrote on 4/27/17 9:37 PM:
>> On Thu, Apr 27, 2017 at 8:28 PM, Shane Curcuru <[email protected]> wrote:
> ...snip...
>> What that means is that your script will have to look at the URL to
>> decide what to output.  Look for "request" in members/watch.cgi for an
>> example.
>
> Fascinating.  Question: when does Wunderbar send data to the browser?
> When I load members/watch/appstatus, it *seems* that the browser renders
> the header and two panels first, then pauses briefly, then renders the
> table.
>
> Which makes me think: shouldn't the txt = File.read(... and txt.scan
> lines be moved after the rendering of the panels, so the user experience
> is better if the server's being slow?

Wunderbar buffers everything up, computes the value of a few header
(etag, for example) and then sends the data all at once.

>> Duplicating sounds bad, but you have plenty of changes, for example
>> linking to the phonebook instead of the roster tool.  There will be
>> some common code, such as parsing the orgchart... that can be factored
>> out to lib/whimsy/asf.
>
> Ah - because while the display of /orgchart and /duties is very similar
> to how it works in roster, much of the rest of the connections to groups
> and roster/committer pages would be completely different than the
> private version.
>
> Yes, separately I wonder sometimes why some code is in this model, vs.
> moving back to whimsy/asf.  Some differences absolutely make sense, but
> on first inspection it feels like there are some other parts that could
> get factored back.  I suppose that's 1) how things got built
> organically, and 2) only has a purpose if some other tool needs those
> functions as well.

Pretty much.

> Starting to feel like I might still be a developer after all!

I've been impressed by how fast you've come on board.

> --
>
> - Shane
>   https://www.apache.org/foundation/marks/resources

- Sam Ruby

Reply via email to