Drop support for windows

[normanmaurer]

I propose to drop support for windows and just focus on linux / macOS. Users of windows can just use the JDK implementation for SSL

[normanmaurer]

/cc @vietj @slandelle @trustin @chrisvest @ejona86 @franz1981

[franz1981]

I am not aware of anything bad with it, but let me summon @cescoffier here

[cescoffier]

+1

[normanmaurer]

/cc @violetagg

[slandelle]

I’m fine with it. What’s less work for you.

[violetagg]

+1

[ejona86]

CC @matthewstevenson88 @rmehta19

[ejona86]

CC @kannanjgithub

[vietj]

@normanmaurer I'm fine with that

[chrisvest]

I'm biased, but I'm in favor. We already don't support any native transports for Windows.

[zekronium]

Isin't the windows JDK implementation much worse performance wise than BoringSSL?

[normanmaurer]

@zekronium maybe ? Honestly I don't know ... That said no-one of our core developers has any use-case for windows so I feel we should better invest our energy in things that matter to us :)

[normanmaurer]

@ejona86 so any concerns from the GRPC side or not ?

[trustin]

I'm fine with dropping Windows support, but I'm curious about why. Is it because of complex build and release process?

[zekronium]

Windows seems least complex to build atleast for me. What is windows specific here that requires/required significant amount of maintenance work?