prusa-slicer: 2.7.4 -> 2.8.0

[Rose-David]

Description of changes

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

---

Add a 👍 reaction to pull requests you find important.

[Rose-David]

I'm trying to build it using nix-build -A prusa-slicer, but it fails with the following errors every time:

/nix/store/2p1sq1nr09xr3xb1a9lrjgdanvk1aakb-binutils-2.42/bin/ld: slic3r/liblibslic3r_gui.a(UserAccountCommunication.cpp.o): undefined reference to symbol 'EVP_DigestInit_ex@@OPENSSL_3.0.0'

/nix/store/2p1sq1nr09xr3xb1a9lrjgdanvk1aakb-binutils-2.42/bin/ld: /nix/store/jv2p6cmx23ihj5y4r98wnn2nmv4qhfh5-openssl-3.0.14/lib/libcrypto.so.3: error adding symbols: DSO missing from command line

I'm not sure what I am doing wrong.

[30350n]

Sounds like it's this: PrusaSlicer #12884

Also, this would fix #322910.

[VuiMuich]

Sounds like it's this: PrusaSlicer #12884

Also, this would fix #322910.

I tried the patch @mogorman posted in the linked upstream issue, but that didn't work for me. The linked gentoo patches though worked. so here they are are for reference:
Missing includes:

--- a/src/slic3r/Config/Version.cpp
+++ b/src/slic3r/Config/Version.cpp
@@ -7,6 +7,7 @@
 #include <cctype>
 
 #include <boost/filesystem/operations.hpp>
+#include <boost/filesystem/directory.hpp>
 #include <boost/nowide/fstream.hpp>
 
 #include "libslic3r/libslic3r.h"
--- a/src/slic3r/GUI/UserAccountCommunication.cpp
+++ b/src/slic3r/GUI/UserAccountCommunication.cpp
@@ -13,6 +13,7 @@
 #include <boost/filesystem.hpp>
 #include <boost/nowide/cstdio.hpp>
 #include <boost/nowide/fstream.hpp>
+#include <boost/nowide/convert.hpp>
 #include <curl/curl.h>
 #include <string>

and fixed linkin:

--- a/src/slic3r/CMakeLists.txt
+++ b/src/slic3r/CMakeLists.txt
@@ -376,6 +376,7 @@ set(SLIC3R_GUI_SOURCES
 )
 
 find_package(NanoSVG REQUIRED)
+find_package(OpenSSL REQUIRED)
 
 if (APPLE)
     list(APPEND SLIC3R_GUI_SOURCES
@@ -404,7 +405,7 @@ endforeach()
 
 encoding_check(libslic3r_gui)
 
-target_link_libraries(libslic3r_gui libslic3r avrdude libcereal imgui libvgcode GLEW::GLEW OpenGL::GL hidapi libcurl ${wxWidgets_LIBRARIES} NanoSVG::nanosvg NanoSVG::nanosvgrast)
+target_link_libraries(libslic3r_gui libslic3r avrdude libcereal imgui libvgcode GLEW::GLEW OpenGL::GL hidapi libcurl ${wxWidgets_LIBRARIES} NanoSVG::nanosvg NanoSVG::nanosvgrast OpenSSL::SSL OpenSSL::Crypto)
 
 if (MSVC)
     target_link_libraries(libslic3r_gui Setupapi.lib)

As a side note: linking used up a lot of memory (only have 8 gigs + 16GB swap) and froze my system once, so I built it with the --cores 3 flag and closed any other memory hungry programs.

Edit: just realized, that the patch from @mogorman is identical with the linking fix only, so I guess that's why it didn't build, becaus it was missing the includes still..

Edit 2: the patched package builds and runs perfectly fine on x86_64-linux

[K900]

The wxgtk branch should also be updated to match latest upstream. Also, we can cherry-pick the patch instead of waiting for it to be merged upstream.

[Rose-David]

I'll go ahead and add the rest of the gentoo patches to my upstream PR, I didn't know if I needed both or just one.

[Rose-David]

The upstream PR should work now.

[K900]

Also I'm not sure if this is actually a good idea but #331539 is actually how I ran into this PR in the first place.

[VuiMuich]

I build it with all the previous and @K900's clang patch (actually only the part below) again using only --cores 3 and it was 1h21m23s whereas the before build took 2h7m55s...

[K900]

What hardware is that on? On my 7950X3D with --cores 32 (default) it takes, like, single digit minutes.

[VuiMuich]

What hardware is that on? On my 7950X3D with --cores 32 (default) it takes, like, single digit minutes.

LENOVO ThinkPad X1 Carbon 2nd - Intel® Core™ i5-4300U CPU @ 1.90GHz (4) - 8GB RAM ☺️

[K900]

I've merged #331539, this will probably need a rebase now.

[Rose-David]

@K900 I synced the fork, should I remove the "FIXME: remove in 2.8.0" code your PR added?

[K900]

Yes. Also, please rebase instead of merging.

[Rose-David]

How do I do that? This is my first attempt at doing something like this, so I am still very new.

[K900]

It depends on how your nixpkgs remote is set up.

[Rose-David]

My nixpkgs is set up on my computer with two remotes, origin (my nixpkgs fork), and upstream (upstream nixpkgs).

[K900]

Then git fetch upstream followed by git rebase -i upstream/master, apply your second commit as a fixup, then git push -f.

[Rose-David]

Did I do it right? Just making sure.

[K900]

Mostly, but please change the commit message to the standard format (prusa-slicer: 2.7.4 -> 2.8.0), we have tooling that relies on that.

[Rose-David]

Okay, if we're all set, should we cherry-pick the patch like you suggested (and how do I do that), or should we wait for it to be merged upstream (might take a while)

[K900]

Look into the patches attribute on derivations and the fetchpatch function.

[Rose-David]

I think I figured it out! I got it to compile, and it runs!

[Rose-David]

I guess that means it's time to make this a normal PR instead of a draft?

[K900]

Feel free to undraft this when you think it's OK to merge.

[Rose-David]

Update: the login feature doesn't work on my build. I tried to update wxwidgets to see if that was the issue, but it doesn't seem to want to compile afterward.

Nevermind, this appears to be an upstream issue. We will probably be waiting a while for that to be fixed, so we should be fine to merge.

[tko]

Result of nixpkgs-review pr 325590 run on aarch64-darwin 1

1 package failed to build:

• prusa-slicer

The error looks to be the same as with 2.7.4 currently so not introduced by this PR:

no template named 'binary_function' in namespace 'boost::functional::detail'; did you mean '__binary_function'?

error: builder for '/nix/store/kg16943n1fm1yap9mh01hvma41ba6q1c-prusa-slicer-2.8.0.drv' failed with exit code 2;
       last 25 log lines:
       > include/boost/functional.hpp:45:24: note: '__unary_function' declared here
       >             using std::unary_function;
       >                        ^
       > include/boost/functional.hpp:487:68: error: no template named 'binary_function' in namespace 'boost::functional::detail'; did you mean '__binary_function'?
       >     class const_mem_fun1_ref_t : public boost::functional::detail::binary_function
       >                                         ~~~~~~~~~~~~~~~~~~~~~~~~~~~^
       > include/boost/functional.hpp:46:24: note: '__binary_function' declared here
       >             using std::binary_function;
       >                        ^
       > include/boost/functional.hpp:533:73: error: no template named 'unary_function' in namespace 'boost::functional::detail'; did you mean '__unary_function'?
       >     class pointer_to_unary_function : public boost::functional::detail::unary_function
       >                                              ~~~~~~~~~~~~~~~~~~~~~~~~~~~^
       > include/boost/functional.hpp:45:24: note: '__unary_function' declared here
       >             using std::unary_function;
       >                        ^
       > include/boost/functional.hpp:557:74: error: no template named 'binary_function' in namespace 'boost::functional::detail'; did you mean '__binary_function'?
       >     class pointer_to_binary_function : public boost::functional::detail::binary_function
       >                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~~^
       > include/boost/functional.hpp:46:24: note: '__binary_function' declared here
       >             using std::binary_function;
       >                        ^
       > 16 errors generated.
       > make[2]: *** [src/libslic3r/CMakeFiles/libslic3r.dir/build.make:77: src/libslic3r/CMakeFiles/libslic3r.dir/cmake_pch.hxx.pch] Error 1
       > make[1]: *** [CMakeFiles/Makefile2:1078: src/libslic3r/CMakeFiles/libslic3r.dir/all] Error 2
       > make: *** [Makefile:146: all] Error 2

[30350n]

Update: the login feature doesn't work on my build. I tried to update wxwidgets to see if that was the issue, but it doesn't seem to want to compile afterward.

Nevermind, this appears to be an upstream issue. We will probably be waiting a while for that to be fixed, so we should be fine to merge.

Could this be related to this?

IMPORTANT note for Linux users

PrusaSlicer now depends on WebKit library, which greatly complicates its distribution. Latest Linux distributions (such as Ubuntu 24.04, Fedora 40) ship with newer version of WebKit than older (but still supported) distros. Bundling WebKit into the AppImage is difficult because it leads to various issues with its dependencies. So far we did not manage to create an AppImage that would reliably run on all relevant distros.

I don't see webkit being added as a dependency anywhere in here.

[Rose-David]

Could this be related to this?

I don't think so, given that people are even experiencing this issue on windows.

I don't see webkit being added as a dependency anywhere in here.

That's my bad, but I tried all three versions of webkitgtk that are available on nixpkgs, and none of them changed the outcome.

There's an invisible "Connection Failed" message, (see my comment on PrusaSlicer Issue 12969) but that's all.

[30350n]

There's an invisible "Connection Failed" message.

Yeah I got that too, but before that I also shortly got a message telling me to install glib-networking. So I added that to the buildInputs and now it's working.

Still not super sure how webkit is being linked exactly, but this seems to do the trick.

[Rose-David]

Still not super sure how webkit is being linked exactly, but this seems to do the trick.

Interesting, I'll try this when I get home and add it to the PR.

[Rose-David]

Sure enough, it works. No idea how its working without webkit, but we take those lol. I'll make sure to forward this fix to the folks having the same problem on the prusa repo.