Enable flook #50

pfebrer · 2023-11-28T19:22:16Z

Checklist

Used a personal fork of the feedstock to propose changes
Bumped the build number (if the version is unchanged)
Reset the build number to 0 (if the version changed)
Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
Ensured the license file is being packaged.

conda-forge-webservices · 2023-11-28T19:22:24Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

zerothi · 2023-11-28T19:23:51Z

build-number update required, otherwise we can check it.

pfebrer · 2023-11-28T19:24:46Z

build-number update required, otherwise we can check it.

You are too fast, I was doing it :)

pfebrer · 2023-11-28T19:25:01Z

@conda-forge-admin, please rerender

pfebrer · 2023-11-28T19:41:40Z

Is there some way to try with the current master of siesta? Or is it only possible with tagged releases?

pfebrer · 2023-11-28T19:44:11Z

The MAC build seems to have gone over flook without problems, only the linux one has failed

zerothi · 2023-11-28T19:44:59Z

Is there some way to try with the current master of siesta? Or is it only possible with tagged releases?

conda releases should stick to a formal release, IMHO. So I think you have these options:

wait until rc2 (or 5.0.0)
try and fix the flook compilation by adding a patch which addresses only this option.

From the looks of it, then it seems the problem will be present in later commits as well. It could be env-vars. Play with that.

pfebrer · 2023-11-28T19:47:13Z

Yes, I just wanted to try here. What you build here in the PR doesn't get published, does it?

zerothi · 2023-11-28T19:47:51Z

No, only once merged. :)

pfebrer · 2023-11-28T19:48:51Z

Do you think this:

gcc: No such file or directory

Means that gcc is not available, or is it an error of gcc not finding some file? 😅

zerothi · 2023-11-28T19:49:45Z

typically conda gcc on linux is named very weirdly, look through the look files for GCC_CC or similar, then you'll find long names. :)

pfebrer · 2023-11-28T23:46:09Z

@conda-forge-admin, please rerender

pfebrer · 2023-11-29T00:30:13Z

Not pretty, but it works :)

zerothi

Yeah, it is too hacky, could you try with fixing the setup.make instead?

zerothi · 2023-11-29T07:31:33Z

recipe/build.sh

@@ -64,13 +64,29 @@ else
  MPI=ON
 fi

+if [[ "$(uname)" == "Linux" ]]; then
+  # Flook compilation on linux needs some hacks


I don't like this. Very non-standard.

I would suggest you fix the setup.make input file instead, via a patch.

This should also ensure that readline is captured correctly.

zerothi · 2023-11-29T07:31:56Z

recipe/meta.yaml

@@ -44,6 +44,7 @@ requirements:
    - git
    - cmake
    - pkg-config
+    - readline


Readline shouldn't be in build section.

pfebrer · 2023-11-29T09:19:38Z

I have the impression that CMake_C_Compiler is simply not set. In MAC, a gcc binary just happens to be in the path, that's why it works.

https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=832491&view=logs&j=65769002-1136-5bbb-5e91-b3fbb7cdd70e&t=fdb86174-6733-5ed6-1dd5-f2594c29f582&l=2456

The CMake_*_compiler are defined here:

https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=832491&view=logs&j=65769002-1136-5bbb-5e91-b3fbb7cdd70e&t=fdb86174-6733-5ed6-1dd5-f2594c29f582&l=491

But only the fortran ones are set. I looked through the CMakeLists file and I can't find where Compilers is requested. Do you know where this is? The right fix would be then to also ask for the C compiler to be defined.

zerothi · 2023-11-29T09:25:00Z

I think you should just patch External/Lua-Engine/setup.make.in, I don't know why C_COMPILER isn't specified in the raw inputs. :(

pfebrer · 2023-11-29T10:04:15Z

Isn't that equally hacky? I think this is a problem for any SIESTA compilation, not just the conda one. It only works if there is by chance a gcc in the path. Perhaps I should submit an issue on the siesta repo instead.

zerothi · 2023-11-29T10:15:23Z

Isn't that equally hacky? I think this is a problem for any SIESTA compilation, not just the conda one. It only works if there is by chance a gcc in the path. Perhaps I should submit an issue on the siesta repo instead.

No, not really. The patch should point the CC to the correct GCC variable. In this case it is how conda-forge setups its build procedure. And that may be very different from local installations.
Moving and copying is executables are much more hacky since they can use other linker lines depending on location (which can be dangerous).

I think this is not a problem in the siesta repo as the cmake invocation correctly sets the CC variable in setup.make, so IMHO it looks like a missing variable in conda.

jan-janssen · 2023-11-29T10:23:25Z

You can use:

cmake -S. -Bobj_cmake "${cmake_opts[@]}" ${CMAKE_ARGS}

For more explanations https://conda-forge.org/docs/maintainer/knowledge_base.html#using-cmake

zerothi · 2023-11-29T10:25:31Z

You can use:
cmake -S. -Bobj_cmake "${cmake_opts[@]}" ${CMAKE_ARGS} 
For more explanations https://conda-forge.org/docs/maintainer/knowledge_base.html#using-cmake

Ah, I always assumed CMAKE_ARGS was automatically read, then please try that here, yes! :)

pfebrer · 2023-11-29T11:31:23Z

Ok, so I found out that CC in setup.make is set to:

$BUILD_PREFIX/bin/x86_64-conda-linux-gnu-cc

which exists, and it is a symbolic link to

$BUILD_PREFIX/bin/x86_64-conda-linux-gnu-gcc

So it is not a problem of cmake not finding the compiler. But somehow this is ignored and gcc is used instead.

pfebrer · 2023-11-29T11:44:55Z

I think flook just ignores whatever you set up as C compiler in setup.make.

I tried with:

PLATFORM=linux
CC = fake_cc
FC = fake_fc
#
.f90.o:
        $(FC) -c $(FFLAGS) $(INC) $<
.F90.o:
        $(FC) -c $(FFLAGS) $(INC) $<
.c.o:
        $(CC) -c $(CFLAGS) $(INC) $<

And I get:

setup.make:9: warning: overriding recipe for target '.f90.o'
setup.make:9: warning: ignoring old recipe for target '.f90.o'
setup.make:11: warning: overriding recipe for target '.F90.o'
setup.make:11: warning: ignoring old recipe for target '.F90.o'
setup.make:13: warning: overriding recipe for target '.c.o'
setup.make:13: warning: ignoring old recipe for target '.c.o'
(mkdir -p aotus/obj ; cd aotus/obj ; \
	echo "TOP_DIR =../.././aotus" > Makefile ; \
	echo "PLATFORM = linux" >> Makefile ; \
	echo "include ../.././aotus/Makefile.smeka" >> Makefile )
(cd aotus/obj ; make SETUP=../../setup.make )
make[1]: Entering directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj'
mkdir -p external
cp -rf ../.././aotus/external/lua-5.3.5 external/
make -C external/lua-5.3.5 linux
make[2]: Entering directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5'
cd src && make linux
make[3]: Entering directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5/src'
make all SYSCFLAGS="-DLUA_USE_LINUX" SYSLIBS="-Wl,-E -ldl -lreadline"
make[4]: Entering directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5/src'
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lapi.o lapi.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lcode.o lcode.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lctype.o lctype.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ldebug.o ldebug.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ldo.o ldo.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ldump.o ldump.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lfunc.o lfunc.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lgc.o lgc.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o llex.o llex.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lmem.o lmem.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lobject.o lobject.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lopcodes.o lopcodes.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lparser.o lparser.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lstate.o lstate.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lstring.o lstring.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ltable.o ltable.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ltm.o ltm.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lundump.o lundump.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lvm.o lvm.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lzio.o lzio.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lauxlib.o lauxlib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lbaselib.o lbaselib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lbitlib.o lbitlib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lcorolib.o lcorolib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ldblib.o ldblib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o liolib.o liolib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lmathlib.o lmathlib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o loslib.o loslib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lstrlib.o lstrlib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o ltablib.o ltablib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lutf8lib.o lutf8lib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o loadlib.o loadlib.c
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o linit.o linit.c
ar rcu liblua.a lapi.o lcode.o lctype.o ldebug.o ldo.o ldump.o lfunc.o lgc.o llex.o lmem.o lobject.o lopcodes.o lparser.o lstate.o lstring.o ltable.o ltm.o lundump.o lvm.o lzio.o lauxlib.o lbaselib.o lbitlib.o lcorolib.o ldblib.o liolib.o lmathlib.o loslib.o lstrlib.o ltablib.o lutf8lib.o loadlib.o linit.o 
ar: `u' modifier ignored since `D' is the default (see `U')
ranlib liblua.a
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o lua.o lua.c
gcc -std=gnu99 -o lua   lua.o liblua.a -lm -Wl,-E -ldl -lreadline 
gcc -std=gnu99 -O2 -Wall -Wextra -DLUA_COMPAT_5_2 -DLUA_USE_LINUX    -c -o luac.o luac.c
gcc -std=gnu99 -o luac   luac.o liblua.a -lm -Wl,-E -ldl -lreadline 
make[4]: Leaving directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5/src'
make[3]: Leaving directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5/src'
make[2]: Leaving directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj/external/lua-5.3.5'
ar -x external/lua-5.3.5/src/liblua.a
fake_fc  -c -o dump_lua_fif_module.o  -Iexternal/lua-5.3.5/src ../.././aotus/LuaFortran/dump_lua_fif_module.f90
make[1]: fake_fc: Command not found
make[1]: *** [../.././aotus/smeka/Makefile.compiler:141: dump_lua_fif_module.o] Error 127
make[1]: Leaving directory '/home/pfebrer/Simulations/siesta/siesta-feedstock/build_artifacts/siesta_1701257065502/work/obj_cmake/External/Lua-Engine/flook-prefix/src/flook-build/aotus/obj'
make: *** [Makefile:67: lib-aotus] Error 2

So it uses gcc to compile C files but then it tries to use FC for fortran, and it fails (as it should because fake_fc doesn't exist).

pfebrer · 2023-11-29T12:17:41Z

It also ignores the includes (I think it just ignores the recipe for .c.o), that's why readline headers are not found.

So I think this is a problem that has to be solved in flook. Here it can only be solved in a hacky manner.

zerothi · 2023-11-29T20:36:00Z

Perhaps a way around this would be to:

cd flook/aotus/external/lua
make CC=$GCC linux

which should respect it, then a subsequent make shouldn't rebuild as it is already done.
Try out locally ;)

pfebrer · 2023-11-30T10:14:05Z

Shouldn't this be managed by flook? It will happen to all users trying to compile SIESTA, not just here in the conda compilation. Also the proposed workaround adds the extra requirement that flook must be pre checked out, it can't be downloaded on the fly by cmake during the compilation.

zerothi · 2023-11-30T10:15:22Z

Shouldn't this be managed by flook? It will happen to all users trying to compile SIESTA, not just here in the conda compilation. Also the proposed workaround adds the extra requirement that flook must be pre checked out, it can't be downloaded on the fly by cmake during the compilation.

Agreed, the problem is that Lua is compiled in aotus, not in flook. So this is a 3-level nested dependency.

Fixed in aotus
Then flook doesn't need any changes.

But I agree that a temporary commit could be done in flook

zerothi · 2023-11-30T10:19:42Z

In flook's Makefile you could call Make on lua before calling it on aotus, no?

I think not (I could be wrong though).
I think it is called through aotus.

pfebrer · 2023-11-30T10:21:08Z

So then, there could be a bash script or something?

zerothi · 2023-11-30T10:22:12Z

I think it is a matter of just cd'ing into aotus/lua and build it manually, then when one does siesta installation, it won't run the compilation again.

pfebrer · 2023-11-30T10:23:48Z

Yes, but I think that is an odd requirement for users.

Just as you cd into aotus here:

https://github.com/ElectronicStructureLibrary/flook/blob/36964ea2770654b32387cc1ed4b5ffd0d92c791f/Makefile#L66-L70

Can't you cd into lua first?

zerothi · 2023-11-30T10:25:24Z

Yes, I am not talking about what we should do in flook, I am talking about what you should do, now, to fix conda build here.

You still want to enable this at the current source tree of siesta+flook, which is hard-coded/fixed. So there is nothing to be done in flook, you have to work around that problem here.

pfebrer · 2023-11-30T10:28:57Z

I can work around the problem by just creating a symbolic link gcc -> real_c_compiler and it is much less complicated :)

If you agree that this is a bug and could be fixed in flook (and it seems like an easy fix), I don't see why we need to make things complicated (and waste time waiting for the very slow build feedback loop).

zerothi · 2023-11-30T10:33:12Z

I can work around the problem by just creating a symbolic link gcc -> real_cc_compiler and it is much less complicated :)

If you agree that this is a bug and could be fixed in flook (and it seems like an easy fix), I don't see why we need to make things complicated (and waste time waiting for the very slow build feedback loop).

I don't like doing symbolic links in conda-builds. I am worried it has side-effects we are not aware of. That is why I suggest you to change these things in this build procedure. Otherwise we should wait until 5.0.0-rc2 is released. Or even a later version.

I don't have time this week (likely not before February) to fix stuff in aotus/flook.

Again, a new fix in flook won't be used in this build as the version is fixed from the 5.0.0-rc1, as it should be.

pfebrer · 2023-11-30T10:39:51Z

I don't like doing symbolic links in conda-builds. I am worried it has side-effects we are not aware of.

The C compiler that cmake uses (and that should be passed to lua make) is already a symbolic link, x86_64-conda-linux-gnu-cc -> x86_64-conda-linux-gnu-gcc.

pfebrer · 2023-11-30T10:42:49Z

I'm afraid if I try to do it in the other way I will spend the whole day looking at logs and figuring out details with one iteration every 5 minutes, and it is very discouraging. But whatever, I can try it again when I have nothing to do.

zerothi · 2023-11-30T11:05:53Z

I think you have somewhat misunderstood what I have tried to convey...

I don't really have much time now, so if you could help with the below things it would be much simpler.

In flook-repo

create a PR that fixes this issue, it should be a minimal intervention if it touches the aotus subdirectory as that is copied directly from the aotus repo
Ensure it picks-up the setup.make variables

In the siesta-feedstock-repo

Apply the same method you did in the PR for flook, but in the build.sh, then you shouldn't be required to do any weird testing as it should have been tested.
Again, this can be tested locally as you can checkout the 0.8.1 release, and try a simple build script to see if it picks-up the correct variables. So all-in-all everything should be testable locally, at least to a large extent.

pfebrer · 2023-11-30T16:13:02Z

@conda-forge-admin, please rerender

pfebrer · 2023-11-30T18:32:10Z

Ok, so I have come up with a Makefile that works. However I don't think it should be used in general because it calls make on aotus/external/lua* instead of aotus/obj/external/lua*.

If I remove readline from build, the build fails because it can not find readline/readline.h.

pfebrer · 2024-02-05T21:10:15Z

This is very confusing, the only build that fails is linux_64_mpinompi, because it doesn't find readline's headers. But if I build locally (using the build_locally.py script) it works. What's the difference between CI and building locally? They both use a Docker container, no?

I need to publish a tutorial for SIESTA using lua, and the best thing would be to just ask people to install from conda. If someone can help here that would be really great, I don't know what else to try :(

pfebrer · 2024-02-06T20:20:14Z

So finally I took the approach of requiring lua as a dependency so that we don't have to compile it. It avoids modifying anything in flook.

The only extra thing to do is to tell cmake to find and link the lua library when compiling SIESTA, because conda installs it as a dynamic library and therefore it can't be included in libflookall (or at least I don't know how).

As you can see it is all green 🟢 ✔️ So @albgar @jan-janssen @zerothi can any of you merge this?

zerothi

Only thing that needs changing is that the cmakelists.txt should be a patch file, not a copy file.

zerothi · 2024-02-14T10:05:16Z

recipe/build.sh

+# include it in libflookall and we need to explicitly tell cmake
+# to find the lua library and link it. That's what we do in this modified
+# CMakeLists.txt that we copy to flook's directory.
+cp $RECIPE_DIR/flook_patches/CMakeLists.txt External/Lua-Engine


Could you change the CMakeLists.txt to a patch, instead of a full file?

See on the 4.1 branch how to add a patch to the meta.yaml file.

zerothi · 2024-02-14T10:58:27Z

And remove the left over file. Then we should just wait for the green lights.

pfebrer · 2024-02-14T17:59:56Z

All green 🟢

zerothi · 2024-02-14T19:22:00Z

thanks!!!

pfebrer · 2024-02-14T19:40:28Z

Now it failed in main due to a timeout 😅

zerothi · 2024-02-14T19:48:12Z

Now it failed in main due to a timeout 😅

annoying, and I can't get it to rerun until all the others have finished, basically the tests for this branch also required two re-runs... :(

pfebrer requested review from albgar, jan-janssen and zerothi as code owners November 28, 2023 19:22

zerothi requested changes Nov 29, 2023

View reviewed changes

pfebrer mentioned this pull request Nov 29, 2023

Rules for compiling C files are ignored for lua compilation ElectronicStructureLibrary/flook#11

Open

pfebrer force-pushed the enable-flook branch from 2a357c3 to d3c1273 Compare November 30, 2023 15:51

pfebrer force-pushed the enable-flook branch from 5eb3791 to 40d4f42 Compare November 30, 2023 17:05

pfebrer force-pushed the enable-flook branch from 14e9637 to 3027a46 Compare February 5, 2024 20:34

pfebrer force-pushed the enable-flook branch 3 times, most recently from a6734e1 to b3becd4 Compare February 6, 2024 19:33

zerothi requested changes Feb 14, 2024

View reviewed changes

Enable flook

5e8e3f9

pfebrer force-pushed the enable-flook branch from ed6970a to 5e8e3f9 Compare February 14, 2024 11:13

zerothi merged commit 431855f into conda-forge:main Feb 14, 2024
8 checks passed

pfebrer mentioned this pull request Feb 14, 2024

Add lua support #35

Closed

Enable flook #50

Enable flook #50

Conversation

pfebrer commented Nov 28, 2023 • edited Loading

conda-forge-webservices bot commented Nov 28, 2023

zerothi commented Nov 28, 2023

pfebrer commented Nov 28, 2023

pfebrer commented Nov 28, 2023

pfebrer commented Nov 28, 2023

pfebrer commented Nov 28, 2023

zerothi commented Nov 28, 2023

pfebrer commented Nov 28, 2023

zerothi commented Nov 28, 2023

pfebrer commented Nov 28, 2023

zerothi commented Nov 28, 2023

pfebrer commented Nov 28, 2023

pfebrer commented Nov 29, 2023

zerothi left a comment

Choose a reason for hiding this comment

zerothi Nov 29, 2023

Choose a reason for hiding this comment

zerothi Nov 29, 2023

Choose a reason for hiding this comment

pfebrer commented Nov 29, 2023

zerothi commented Nov 29, 2023

pfebrer commented Nov 29, 2023

zerothi commented Nov 29, 2023

jan-janssen commented Nov 29, 2023

zerothi commented Nov 29, 2023 • edited by jan-janssen Loading

pfebrer commented Nov 29, 2023 • edited Loading

pfebrer commented Nov 29, 2023

pfebrer commented Nov 29, 2023

zerothi commented Nov 29, 2023

pfebrer commented Nov 30, 2023 • edited Loading

zerothi commented Nov 30, 2023

zerothi commented Nov 30, 2023

pfebrer commented Nov 30, 2023

zerothi commented Nov 30, 2023

pfebrer commented Nov 30, 2023

zerothi commented Nov 30, 2023

pfebrer commented Nov 30, 2023 • edited Loading

zerothi commented Nov 30, 2023

pfebrer commented Nov 30, 2023

pfebrer commented Nov 30, 2023

zerothi commented Nov 30, 2023

pfebrer commented Nov 30, 2023

pfebrer commented Nov 30, 2023

pfebrer commented Feb 5, 2024

pfebrer commented Feb 6, 2024 • edited Loading

zerothi left a comment

Choose a reason for hiding this comment

zerothi Feb 14, 2024

Choose a reason for hiding this comment

zerothi Feb 14, 2024

Choose a reason for hiding this comment

zerothi commented Feb 14, 2024

pfebrer commented Feb 14, 2024

zerothi commented Feb 14, 2024

pfebrer commented Feb 14, 2024

zerothi commented Feb 14, 2024

pfebrer commented Nov 28, 2023 •

edited

Loading

zerothi commented Nov 29, 2023 •

edited by jan-janssen

Loading

pfebrer commented Nov 29, 2023 •

edited

Loading

pfebrer commented Nov 30, 2023 •

edited

Loading

pfebrer commented Nov 30, 2023 •

edited

Loading

pfebrer commented Feb 6, 2024 •

edited

Loading