<jonpsy[m]>
<rcurtin[m]> "jonpsy: I've been climbing for a..." <- Wait, zoq you climb too? Cool stuff! I was actually thinking of trying out rock climbing. Adding it to my wish list after Horse riding, bungee jumping & surfing :)
<zoq[m]>
<jonpsy[m]> "Wait, zoq you climb too? Cool..." <- I do, not as active anymore that I used to be though; so Ryan is definitely in better shape.
<jonpsy[m]>
interesting, what physical activity are you upto these days? Btw Have you thought of learning horse riding ?
<ShubhamAgrawal[m>
zoq: Is there anything else left in this PR? Except for migrating CLI changes and resolving conflicts?
<rcurtin[m]>
jonpsy: yeah, I highly recommend rock climbing; if you have a nearby gym, it can be a nice social activity too. For me it has provided a lot of stress relief over the years (even if my fingers pay the price this weekend... still not 100% yet 😃)
<rcurtin[m]>
also, does anyone have time to review https://github.com/mlpack/mlpack/pull/3291 today? once that one is merged, I can open the 4.0.0 release PR, and maybe we can have that released tomorrow 😃
<zoq[m]>
<rcurtin[m]> "also, does anyone have time to..." <- Yes, can approve it in a few minutes, already did the review.
<rcurtin[m]>
thanks!
<rcurtin[m]>
if someone is willing to give that a second review, we can merge it today and I can basically open the release PR immediately
<rcurtin[m]>
(but I can't do the release PR until that one is merged)
<shrit[m]>
rcurtin: I can not see this one, it says that it is a binary even so it has C++ extension
<rcurtin[m]>
that is strange, I think this is because the line endings changed to DOS line endings, not UNIX ones. I did that because it makes more sense to use DOS line endings for a Windows-specific example, but I guess it makes the diff very hard...
<EshaanAgarwal[m]>
jonpsy: zoq @marcusedel:matrix.org: I am traveling ! Can we shift the meet pls ?
<zoq[m]>
EshaanAgarwal[m]: To?
<EshaanAgarwal[m]>
zoq[m]: Tomorrow works for me ! Same time
<zoq[m]>
EshaanAgarwal[m]: Have to check, if you know you are going to travel, please let us know on advance.
<zoq[m]>
zoq[m]: Do you have an idea why the test is not converging? Also did you debug it, went through the code?
<EshaanAgarwal[m]>
zoq[m]: I apologise. had to go last minute. Maybe you could suggest time ? I can work that out.
<EshaanAgarwal[m]>
zoq[m]: I am not sure ! I went through the code. And made some changes according to that but still the issue persists. I have already pushed my code for that.
<zoq[m]>
EshaanAgarwal[m]: I see, you pushed it yesterday, interested to know if you stepped through the code and checked if the output is reasonable?
<jonpsy[m]>
<rcurtin[m]> "also, does anyone have time to..." <- looks like its merged
<rcurtin[m]>
yep, thanks everyone! doing the release PR now
<EshaanAgarwal[m]>
<zoq[m]> "I see, you pushed it yesterday..." <- Output looked reasonable according to me. Actually I also tried other replay methods in that environment. Hindsight was performing somewhat similar to RandomReplay. I couldn't dig more from then. Please give me some time to return with more results.
<jonpsy[m]>
Funnily, I was just about to comment a change but it was merged that very second 😅
<EshaanAgarwal[m]>
EshaanAgarwal[m]: Meanwhile can you pls review the environment and replay ? I think I might have missed some theoretical point in it.
<jonpsy[m]>
jonpsy[m]: But it was pretty minor change, I thought mentioning `C:/` might not be necessary and we could've said "working directory" unless I'm missing something
<EshaanAgarwal[m]>
EshaanAgarwal[m]: Especially the environment so that I know that at least I got that right. It looked okay to me😅
<jonpsy[m]>
I can try building mlpack 4.0 on `mac m1` if that helps? rcurtin
<rcurtin[m]>
sure, feel free! I think that it should work just fine but always good to test. we don't have an M1 system in our build farm 😀
<jonpsy[m]>
I used to know how to build it on Windows too. That was eons ago 🥲
<jonpsy[m]>
so I'll test with the `master` branch, right? And follow build instructions in our main github page?
<rcurtin[m]>
yeah---it should I imagine be super easy, just install dependencies then cmake and make and run the tests. at least I hope it will be super easy 😀
<jonpsy[m]>
Installing dependencies was a breeze. Started compiling
<jonpsy[m]>
hmm weird
<jonpsy[m]>
compilation was succesful, it's unable to read `#include <mlpack.hpp>` but I've defo `sudo make install ` in the build folder
<jonpsy[m]>
s/compilation/build/
<jonpsy[m]>
and its not building `mlpack_test` so even though `DBUILD_TEST` is on. So `ctest` is failing as well
<rcurtin[m]>
what happens if you do `make test` explicitly? or `make mlpack_test`?
<jonpsy[m]>
now its compiling something extra
<jonpsy[m]>
ig thats mlpack test. Weird that it didn't compile it with `make`
<rcurtin[m]>
yeah, we didn't want it to be a part of `all`, because it would then cause everyone to always build the tests, even if they just wanted to install
<rcurtin[m]>
however, I am surprised it doesn't auto-build it before `ctest`
<jonpsy[m]>
rcurtin[m]: I see. But I did set the flag
<jonpsy[m]>
maybe my `ccmake ` or cache related issue idk
<rcurtin[m]>
right, maybe there is some CMake configuration option that needs to happen... I'm not in a place I can look into it today, but I can check tomorrow, unless you want to dig in and play with CMake or something 😃
<jonpsy[m]>
Wow! Compiling does take a lot out of battery