[gopher] Gopher++ scrapped & Internet Archive -style thingy

[Kim Holviala]

On 20.4.2010 15:40, Mike Hebel wrote:

>> Whatever I've written about the gopher++ extra headers can now be
>> considered as "obsolete". I found a few live sites which just cannot
>> accept anything else than a selector<CRLF> so there's no way I can
>> insert extra headers without breaking stuff.
>>
>> All code regarding the header extensions has been scrapped and
>> deleted, it's all gone for good.
>
> I'm confused here. Is this the client side of things or the server side?

Both. There were servers which didn't want to work with my extra 
headers. This would have required detection mechanisms in the clients 
(do we send teh headers or not?) which would have made clients too 
complicated.

And if the clients cannot be modified to send extra stuff easily, what's 
the point of having extras on the server side?

The only way I found that worked was my original suggestion with no 
extra data on the first request line but still extra header lines after 
it. This was (rightfully) shot down by people here because it requires 
careful timing on the server side and possibly knowledge of the TCP 
packet sizes on the client side.

> If the goal is to keep Gopher moving forward then why not create a
> better server with an expanded protocol? And if it's just your servers
> that do the gopher++ dance then why does it matter if other servers
> don't?

I think gopher is pretty much on life support now :-) and any change 
that breaks even one of the 150 or so servers still running is not going 
into my code.

Actually, I lied a bit before. I keep very good backups so no code was 
lost... I simply got frustrated and deleted all gopher++ traces from my 
_current development version_. If I ever figure out a way to please 
everybody, even the perl one-liner servers that seem to be out there, 
then the code will find it's way back.


- Kim