[tor-dev] Is it time to drop support for the v1/v2 protos?

[grarpamp]

On Fri, Jan 2, 2015 at 12:39 PM, Tom van der Woerdt <info at tvdw.eu> wrote:
>>    Now, maybe we _should_ drop support for versions before
>> 0.2.3.17-beta as well.  If so, we can rip out even more code.  (And
>> that might be a good idea.) What do people on the list think?

X Fabio:
X The cleaner, the better!

> I'd definitely like ripping out more legacy support code :-) According to
> the consensus, 0.2.3.24-rc is the lowest recommended version right now, and
> I (mis)interpreted that as the lowest version that must still be supported
> after I rip out code.
> ...
> Also, in case anyone is wondering why a complete stranger is suddenly so
> interested in ripping parts out of Tor: I attempted to implement a Tor
> client and was constantly faced with legacy stuff in the spec. Might as well
> get rid of it if you don't need it - also makes my client implementation
> easier.

Look at the rend-spec.txt cases of ancient protocols alone. Yes,
I'd try to conform those docs along with this too.

I think legacy protocol support ties both a project and it's users down.
Don't be afraid to rip things out. What is the worst that could happen?
A user has to go install the current release? Oh my, whatever.

There's also a distinct difference between backwards compatible legacy
protocol support (a ton of complex code), and support of older/wider platforms
on the current release (a few small things, support for those is
pretty good btw).

The only time I'd worry about old protocol regarding Tor is if there
is truly a significant number of users for which the old *protocol*
offers them better, or their only, means to anonymity, circumvention,
obfuscation (as opposed to old *porting*). I'm guessing the answer
is probably no here.

The consensus version is what tor thinks is best acceptable minimum
regarding features, security, bugs, anonymity, etc. It is a nag, not a hard.
Writing new client to match anything older than whatever specifications
that version supports would waste your time, especially since even that
will slide by the time you're done writing and testing.