--- Log opened Fri Jun 24 00:00:01 2022
06:00 -!- halosghost [~halosghos@user/halosghost] has joined #secp256k1
06:00 < nickler> Hey. I have no experience with cmake or how it would simplify our life. The advantages you mention in the PR don't seem to apply to libsecp and autotools kind of works right now.
06:00 < nickler> If Core does this, is it difficult to stick with autotools? I suppose we'd have to maintain two build systems in parallel at least for a while to give library users some time to move away from autotools.
06:24 -!- sgiath- [~sgiath@2a02:25b0:aaaa:aaaa:a3c3:ed4b:6b06:0] has quit []
06:24 -!- sgiath [~sgiath@mail.sgiath.dev] has joined #secp256k1
08:37 < hebasto> during migration to CMake in Bitcoin Core, the actual libsecp's build system does not really matter :) it is just been overridden -- https://github.com/hebasto/bitcoin/blob/cmake/pr/cmake/subtree-secp256k1.cmake
08:38 < hebasto> other benefits, not clear highlighted, are ability to build on Windows with MSVC
08:42 < hebasto> another advantage is a larger number of CMake-aware developers; with time, it will get stronger
08:52 < nickler> Ah, nice that it can be so easily overriden. But why are you specifically interested in cmake for libsecp then?
08:54 < hebasto> not specifically :) just suggesting my time if you find  CMake useful for libsecp
08:55 < hebasto> anyway, one could build libsecp with no build system at all :)
09:00 < nickler> sure
09:01 < nickler> thanks for offering your time. The main concern was always that we don't want to maintain two build systems and there was never enough reason to switch.
09:02 < nickler> There's an intermediate solution where the cmake build system lives in a separate repo (e.g. discussed here https://github.com/bitcoin-core/secp256k1/pull/549)
09:03 < sipa> Bitcoin Core (possibly) switching to cmake would be an extra reason to have a cmake build system in libsecp256k1, if there is going to be maintenance attention to it anyway.
09:04 < sipa> Maintaining two build systems is an extra cost though, but I suspect that for the most part, our build system would be pretty simple - and with some of the (slowly...) ongoing refactors it may become even easier.
09:06 < hebasto> btw, there are many CMake configurations for libsecp around, e.g. -- https://github.com/microsoft/vcpkg/blob/master/ports/secp256k1/CMakeLists.txt
09:46 -!- yakshaver123 [yakshaver@2600:3c00::f03c:92ff:fe8e:dce6] has joined #secp256k1
09:47 -!- yakshaver [yakshaver@2600:3c00::f03c:92ff:fe8e:dce6] has quit [Ping timeout: 248 seconds]
11:09 < berryyoghurt> hebasto: Could you please explain how to build libsecp without a build system? I am writing some cross tests (https://github.com/bitcoin-core/secp256k1/pull/1112) which I want to try out. But to try them I need to link another library (libecc) as well as define a macro. I would like to do that with as little hassle as possible (I don't want to concern myself with Autotools right now).. I presume building libsecp without a 
11:09 < berryyoghurt> build system will be a quick way to go?
11:10 < sipa> you can always do by hand exactly the same thing as the build system is doing
11:10 < sipa> if you run make V=1 it'll print the commands
11:38 < hebasto> berryyoghurt: re "Could you please explain how to build libsecp without a build system? " -- https://github.com/bitcoin-core/secp256k1/issues/929
13:46 -!- Netsplit *.net <-> *.split quits: berryyoghurt
13:46 -!- Netsplit over, joins: berryyoghurt
14:06 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Read error: Connection reset by peer]
14:41 -!- halosghost [~halosghos@user/halosghost] has quit [Quit: WeeChat 3.5]
15:33 -!- luke-jr [~luke-jr@user/luke-jr] has joined #secp256k1
15:45 -!- ghost43 [~ghost43@gateway/tor-sasl/ghost43] has quit [Remote host closed the connection]
15:45 -!- ghost43 [~ghost43@gateway/tor-sasl/ghost43] has joined #secp256k1
17:34 -!- BlueMatt [~BlueMatt@ircb.bluematt.me] has quit [Ping timeout: 244 seconds]
17:39 -!- BlueMatt [~BlueMatt@ircb.bluematt.me] has joined #secp256k1
18:25 -!- BlueMatt [~BlueMatt@ircb.bluematt.me] has quit [Quit: Quit]
18:25 -!- BlueMatt [~BlueMatt@ircb.bluematt.me] has joined #secp256k1
--- Log closed Sat Jun 25 00:00:02 2022