prusa-slicer: 2.8.0 -> 2.9.0 #367376

me-and · 2024-12-22T13:44:35Z

Update Prusa Slicer to the latest v2.9.0. This builds on @cbat98's as-yet unmerged PR for v2.8.1 in #358215.

Changelog:

Things done

Add a 👍 reaction to pull requests you find important.

cbat98 · 2024-12-22T15:02:53Z

Thank you for taking a look at the 2.9.0 update :) Appreciate you building on my attempt at 2.8.1.

jmickelin · 2024-12-23T12:55:34Z

Nice stuff! I tested it and it works 👍

@maxammann raised a point in the previous PR that PrusaSlicer exhibits a bug with downloading for Curl versions newer than 7.88.1. Seeing how it's already broken in the current 2.8.0 package in nixpkgs, I think we can leave it as-is for this PR and then discuss overriding the Curl dependency in a followup PR.

me-and · 2024-12-23T17:24:53Z

I am getting a breakage in the super-slicer-beta package, which I've not yet debugged further, or even checked if it exists prior to this change...

CMake Error at src/occt_wrapper/CMakeLists.txt:22 (find_package):
  Could not find a configuration file for package "OpenCASCADE" that is
  compatible with requested version "7.6.2".

  The following configuration files were considered but not accepted:

    /nix/store/yqk4xxvw5pmjjq3ah4dahka5i8yxw0z6-opencascade-occt-7.6.1/lib/cmake/opencascade/OpenCASCADEConfig.cmake, version: 7.6.1



-- Configuring incomplete, errors occurred!

jmickelin · 2024-12-23T19:45:09Z

The issue comes from the following line: https://github.com/supermerill/SuperSlicer/blob/master/src/occt_wrapper/CMakeLists.txt#L22

It's really strange that the super-slicer derivation is coupled to the prusa-slicer one like this, since the former is lagging behind a lot still. They really should be independent packages with their own build requirements these days.

In any case, since super-slicer hardcodes the dependency on OCCT-7.6.2 and prusa-slicer pins it to 7.6.1, I think you need to override it "back" in super-slicer.nix, something like this:

--- a/super-slicer.nix
+++ b/super-slicer.nix  
@@ -6,6 +6,7 @@
   wxGTK31,
   prusa-slicer,
   libspnav,
+  opencascade-occt_7_6
 }:
 let
   appname = "SuperSlicer";
@@ -129,7 +130,10 @@ let
       fetchSubmodules = true;
     };
   });
-  prusa-slicer-wxGTK-override = prusa-slicer.override { wxGTK-override = wxGTK31-prusa; };
+  prusa-slicer-dependencies-override = prusa-slicer.override { 
+    wxGTK-override = wxGTK31-prusa; 
+    opencascade-occt_7_6_1 = opencascade-occt_7_6;
+  };
   allVersions = builtins.mapAttrs (
-    _name: version: (prusa-slicer-wxGTK-override.overrideAttrs (override version))
+    _name: version: (prusa-slicer-dependencies-override.overrideAttrs (override version))
   ) versions;

It looks a bit hacky, with the argument opencascade-occt_7_6_1 being set to (effectively) opencascade-occt-7_6_2 but I think it's inevitable when it's coupled like this. The other option would be to patch SuperSlicer to downgrade the required version. Surely it would encounter the same bug that caused PrusaSlicer to pin it in the first place, so it could make sense: prusa3d/PrusaSlicer@c6a0210#diff-54d1bb7ea7d9b83fb3c587bc042046732a50bb843a067cd86c445bafbb6ba1eeR7

cbat98 · 2024-12-30T18:58:58Z

I don't know if it's something wacky with my setup, but to get it to build for me, I had to also apply this patch:

diff --git a/src/slic3r-arrange/include/arrange/DataStoreTraits.hpp b/src/slic3r-arrange/include/arrange/DataStoreTraits.hpp
index ba932d57a..624569d3f 100644
--- a/src/slic3r-arrange/include/arrange/DataStoreTraits.hpp
+++ b/src/slic3r-arrange/include/arrange/DataStoreTraits.hpp
@@ -71,7 +71,7 @@ template<class T, class TT = T> using WritableDataStoreOnly = std::enable_if_t<I
 template<class T, class ArrItem>
 void set_data(ArrItem &itm, const std::string &key, T &&data)
 {
-    WritableDataStoreTraits<ArrItem>::template set(itm, key, std::forward<T>(data));
+    WritableDataStoreTraits<ArrItem>::template set<T>(itm, key, std::forward<T>(data));
 }

template constexpr bool IsReadWritableDataStore = IsDataStore && IsWritableDataStore;

I'm not sure why this was needed, but I was seeing an error in the build at about 55%.

me-and · 2024-12-31T22:32:43Z

I tried to split out the super-slicer package so it no longer had the dependency on prusa-slicer, but I couldn't work out how to get that working. I think the complicating factor is the use of darwin.apple_sdk_11_0.callPackage in all-packages.nix, but in honesty I didn't care enough to spend a long time on it.

In any case, @jmickelin's approach worked for me, although I modified it slightly to emulate the existing override of wxGTK in super-slicer. I'll push that patch shortly.

@cbat98 I'm baffled that you're seeing that error when it's working fine for me building out of the nixpkgs tree. Possibly you have an overlay confusing things? How do you get on with nix-shell -p nixpkgs-review --run 'nixpkgs-review pr 367376', which I think enforces a clean environment?

me-and · 2024-12-31T23:36:31Z

...oh, that's odd. I'm now seeing a build failure as well, which looks like the same build failure as @cbat98 saw. I was building at the tip of my branch, but it looks like the build is broken by something else that's changed since I branched off...

`nixpkgs-review` result

Generated using nixpkgs-review.

Command: nixpkgs-review

`x86_64-linux`

❌ 2 packages failed to build:

prusa-slicer
prusa-slicer.debug

✅ 5 packages built:

opencascade-occt_7_6_1
super-slicer (super-slicer-latest)
super-slicer-beta
super-slicer-beta.debug
super-slicer.debug (super-slicer-latest.debug)

me-and · 2025-01-03T18:33:51Z

...that was daft. I backed out 910f66b because I'd interpreted the error I saw to mean the patch wasn't necessary. Turns out it just means the file the patch was applying to was in a different file.

Assuming it doesn't find any other problems, I'll push an updated branch as soon as my local build has finished.

At the same time fix the paths in a manual patch, and fix the version of Boost to meet the build requirements.

When overriding the postPatch attribute from prusa-slicer, handle the scenario where prusa-slicer doesn't have a postPatch attribute, as well as the scenario where it does.

super-slicer is built by overriding prusa-slicer, and the updated libraries as part of the update of prusa-slicer to v2.9.0 broke super-slicer's pinned libraries. Ideally we'd split off super-slicer as a separate package rather than continuing to make changes in prusa-slicer then undo them in super-slicer, but that seems to be a much more substantial job, so stick with the simple approach for now.

me-and · 2025-01-03T19:18:00Z

There we go!

`nixpkgs-review` result

Generated using nixpkgs-review.

Command: nixpkgs-review

`x86_64-linux`

✅ 7 packages built:

opencascade-occt_7_6_1
prusa-slicer
prusa-slicer.debug
super-slicer (super-slicer-latest)
super-slicer-beta
super-slicer-beta.debug
super-slicer.debug (super-slicer-latest.debug)

github-actions bot added 10.rebuild-darwin: 1-10 10.rebuild-linux: 1-10 labels Dec 22, 2024

me-and force-pushed the prusa-slicer-tip branch 2 times, most recently from 49a7f8c to b2304c2 Compare December 22, 2024 14:59

me-and force-pushed the prusa-slicer-tip branch from b2304c2 to 52cf5f5 Compare December 22, 2024 20:30

me-and marked this pull request as ready for review December 22, 2024 20:33

me-and added 8.has: package (update) This PR updates a package to a newer version 8.has: package (new) This PR adds a new package labels Dec 22, 2024

ofborg bot requested review from thorstenweber83, gebner, hesiod, cab404 and amiloradovsky December 23, 2024 09:01

me-and marked this pull request as draft December 31, 2024 23:37

cbat98 and others added 8 commits January 3, 2025 19:08

prusa-slicer: Update version and hash

65f18db

prusa-slicer: Set opencascade version

2e23ccf

prusa-slicer: Update dependency with version

beb95e6

prusa-slicer: Remove existing patches

5056f66

Add override for occt 7.6.1

3e94f46

prusa-slicer: Use new overriden occt vn

29ddfad

prusa-slicer: Move boost log link from patch to args

5dfc53b

prusa-slicer: 2.8.1 -> 2.9.0

92d5a75

At the same time fix the paths in a manual patch, and fix the version of Boost to meet the build requirements.

me-and added 2 commits January 3, 2025 19:09

super-slicer: handle missing postPatch

763f17c

When overriding the postPatch attribute from prusa-slicer, handle the scenario where prusa-slicer doesn't have a postPatch attribute, as well as the scenario where it does.

me-and force-pushed the prusa-slicer-tip branch from b432eed to 1bd956f Compare January 3, 2025 19:17

me-and marked this pull request as ready for review January 3, 2025 19:18

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

prusa-slicer: 2.8.0 -> 2.9.0 #367376

prusa-slicer: 2.8.0 -> 2.9.0 #367376

me-and commented Dec 22, 2024 •

edited

Loading

cbat98 commented Dec 22, 2024

jmickelin commented Dec 23, 2024

me-and commented Dec 23, 2024

jmickelin commented Dec 23, 2024 •

edited

Loading

cbat98 commented Dec 30, 2024 •

edited

Loading

me-and commented Dec 31, 2024

me-and commented Dec 31, 2024

me-and commented Jan 3, 2025

me-and commented Jan 3, 2025

prusa-slicer: 2.8.0 -> 2.9.0 #367376

Are you sure you want to change the base?

prusa-slicer: 2.8.0 -> 2.9.0 #367376

Conversation

me-and commented Dec 22, 2024 • edited Loading

Things done

cbat98 commented Dec 22, 2024

jmickelin commented Dec 23, 2024

me-and commented Dec 23, 2024

jmickelin commented Dec 23, 2024 • edited Loading

cbat98 commented Dec 30, 2024 • edited Loading

me-and commented Dec 31, 2024

me-and commented Dec 31, 2024

nixpkgs-review result

x86_64-linux

me-and commented Jan 3, 2025

me-and commented Jan 3, 2025

nixpkgs-review result

x86_64-linux

me-and commented Dec 22, 2024 •

edited

Loading

jmickelin commented Dec 23, 2024 •

edited

Loading

cbat98 commented Dec 30, 2024 •

edited

Loading

`nixpkgs-review` result

`x86_64-linux`

`nixpkgs-review` result

`x86_64-linux`