This week in FreeCAD development:

Sketcher

  • theo-vt pre­vent­ed self-tan­gency from hap­pen­ing at the com­mand lev­el (PR#31042).
  • YashSuthar983 patched sketch visu­al­iza­tion to have two con­fig­urable para­me­ters: trans­paren­cy of the over­all axis and occlud­ed trans­paren­cy for the occlud­ed axis, so that an occlud­ed axis can be seen (PR#26594).
  • theosib fixed auto-con­strain­ing to exter­nal geom­e­try in cer­tain sce­nar­ios (PR#30970).

Assembly

  • LubuSeb changed Assem­bly acti­va­tion to use the same scoped recom­pute path used by the man­u­al Solve Assem­bly com­mand, instead of call­ing the solver direct­ly from edit-mode acti­va­tion (PR#30744). This fix­es a bug where acti­vat­ing an assem­bly would block auto­mat­ic recom­pu­ta­tion after chang­ing the VarSet parameters.
  • Pad­dle­Stroke fixed incor­rect­ly placed new­ly cre­at­ed joints (PR#30669).
  • Lgt2x switched the Assem­bly work­bench from OpenCV to PyAV to reduce the size of the builds with­out any func­tion­al­i­ty loss (PR#30333).

TechDraw

Thanks to pavl­tom, now you can attach any expres­sion sup­port­ed by the FreeCAD expres­sion edi­tor to any of the Tech­Draw tem­plate’s fields (PR#27117).

CAM

  • tarman3 fixed mul­ti-tar­get drilling in Helix (PR#30611) and patched recom­putes in Select­Loop (PR#30908).
  • jffmichi removed unneed­ed recom­putes that would hap­pen every time a sim­u­la­tion was start­ed (PR#30927).
  • dbtayl fixed path gen­er­a­tion fail­ures in Adap­tive (PR#30623).
  • davidg­ilka­uf­man fixed a bug where Adap­tive clear­ing would go around the work­piece perime­ter when “Fin­ish­ing pro­file” is select­ed (PR#30831).

FEM

  • marioalexis84 set the default pres­sure val­ue to 1 MPa (PR#31036).
  • wmay­er imple­ment­ed meth­ods to get ele­ment col­ors (PR#30726, cher­ry-picked by maxwxyz).

Other changes

  • Roy_043 added sup­port for mul­ti­ple objects in hatch­es in Draft (PR#30578).
  • tri­tao added view rota­tion by drag­ging the Nav­iCube (PR#30727).
  • vsergeev added sup­port for ver­ti­cal con­fig­u­ra­tion tables in Spread­sheet (PR#29656).

Addi­tion­al improve­ments and fix­es were con­tributed by Amine0x0, awgrover, chennes, eesh­sax­e­na, greg19, Lgt2x, Roy-043, tarman3, theosib, tri­tao, and wmayer.

If you are inter­est­ed in test­ing the lat­est week­ly build, you can grab it here.

Activity stats

PR stats: since the pre­vi­ous report, 41 pull requests have been merged, and 49 new pull requests have been opened.

Issue stats: over­all, there are 3495 open issues in the track­er, up by 17 from last week.

Elsewhere in the community

Ephi Blan­shey recent­ly shared his progress work­ing on the His­to­ry work­bench that helps ver­sion-con­trol changes in FreeCAD projects native­ly and visu­al­ize dif­fer­ences between mod­el revisions.


Discover more from FreeCAD News

Subscribe to get the latest posts sent to your email.

6 responses to “WIP Wednesday, 1 July 2026”

  1. barbara Avatar
    barbara

    The his­to­ry work­bench fills a need I’ve had for a long time. I am a bit sad­dened to see that it’s vibecod­ed, how­ev­er. Of course, not a big deal for a proof of con­cept, but with that in mind, is there any poten­tial plans to devel­op a built-in ver­sion con­trol func­tion­al­i­ty? Per­haps inspired by said work­bench, and tak­ing into avvount lessons learned in terms of UI design and ergonom­ics. But clean-room designed to avoid the ethics issues that comes with the AI cod­ed one.

    1. douyarou Avatar

      I’m hav­ing trou­ble fol­low­ing your anti-AI slant here. A solo devel­op­er decides to use tools avail­able to them to make some­thing nice for every­one. Gives it away for free, and asks for noth­ing in return, and you’re “sad­dened” and hop­ing some­one else steals their idea and builds it with­out AI? When mega-corps like Microsoft or Google lever­age their enor­mous cap­i­tal to do this, we do not call it eth­i­cal. Depend­ing on the mod­el used here, the com­pa­ny that trained the mod­el may have act­ed uneth­i­cal­ly in the past and/or present, but that nuance is com­plete­ly absent in your com­ment. An end user try­ing to improve oth­ers’ lives does­n’t seem uneth­i­cal to me at all. How about we fix copy­right law first, then apply it equal­ly to every­one, and even pun­ish the ones that broke even the old law, and per­haps we then start build­ing more eth­i­cal AI? Sounds great to me, but since we’re all pre­tend­ing Ope­nAI and the rest did­n’t steal their train­ing mate­r­i­al, let’s not crap on the peo­ple who use AI to learn, build, share, or even try to improve their own mis­er­able lives.

    2. douyarou Avatar

      I don’t see any men­tion of this being vibe cod­ed any­where. On what basis do you attack this work­bench for being vibe cod­ed? Got any links or are you just going off of vibes?

  2. barbara Avatar
    barbara

    Here you go: https://github.com/eblanshey/HistoryWorkbench/blob/b78e2d97db8567f08fd07df88277227036427c7d/AGENTS.md

    First off, I had­n’t found the updat­ed FreeCAD pol­i­cy on LLM use until just after mak­ing my post, so that’s on me. How­ev­er, I fol­low some PRs and issues over on the Github page and accord­ing to what was said there, there was pre­vi­ous a blan­ket ban on all LLM use. So that gave me the impres­sion that a project like this could­n’t ever be merged as-is, which is why I brought up reim­ple­men­ta­tion. I think VCS is so cen­tral to a ful­ly fea­tured CAD suite that it should­n’t have to be done by a plugin.

    With the updat­ed pol­i­cy and the “you’re liable for the code you gen­er­ate” clause, that might not be nec­es­sary. Though I would be hes­i­tant to merge any­thing of this scale (this repo is fair­ly large for a plu­g­in) into the main project, even if it was cre­at­ed by tra­di­tion­al means.

    You have a lot of points, some address­ing things I nev­er said. At least I don’t think I did? Since this is a blog post com­ment chain and not a prop­er forum dis­cus­sion, I don’t see the point of mak­ing it into a huge back and forth debate. But if you real­ly want to engage with the dis­cus­sion and where I am com­ing from, keep reading.

    I nev­er said any­thing about “steal­ing ideas”, in fact, I praised the project and sug­gest­ed that it should be incor­po­rat­ed in the main pro­gram as a clean-room reim­ple­men­ta­tion exact­ly to avoid those kinds of con­cerns. The same way that open source devel­op­ment has always built upon and improved upon uneth­i­cal closed source ideas. And the assump­tion is that this should be done in col­lab­o­ra­tion with the orig­i­nal uploader, and cred­it­ing their work. Like aca­d­e­m­ic endeav­ors have always been done. That’s why I referred to the third par­ty project as a proof-of-concept…

    I don’t under­stand where you got the idea that I am excus­ing the uneth­i­cal behav­ior of the mega­corps that trained the mod­els? Quite the oppo­site? It’s the core of almost all anti-genAI dis­cus­sions, and one of the big two rea­sons why peo­ple are boy­cotting things using it. This has been dis­cussed to death in the tech com­mu­ni­ty, so I did­n’t pep­per my post with caveats to avoid step­ping on bil­lion­aire toes.

    Maybe we fre­quent dif­fer­ent com­mu­ni­ties, but this sen­ti­ment isn’t in any way rad­i­cal. And yes, the ram­pant accep­tance of vib­ing in the open source world does sad­den me. A whole lot. It mir­rors the exact same trend from the tech indus­try that has lost both me and my close friends their liveli­hood. And puts fuel on the Bran­dolin­i’s law-shaped fire that threat­ens my new career as a teacher.

    You’re free to make your case that there will some day be an eth­i­cal AI (I dis­agree), but even you agree that we aren’t there yet.

    One of the things we as open source con­trib­u­tors have, that indus­try projects do not, is time. Proof-of-con­cepts and licens­ing workarounds cre­ate tech debt. But unlike for-prof­it shops, there is both plen­ty of time and vol­un­teers to turn them into unprob­lem­at­ic, clean, fea­tures that can be main­lined. Obvi­ous­ly the ball would­n’t roll any­where with­out any­one will­ing to make a PoC. Now they have, so we should­n’t be stuck argu­ing about how it can nev­er be improved upon, and dig­ging our heels into the ground for ide­o­log­i­cal rea­sons when the sug­ges­tion comes up.

    Bet­ter is bet­ter. Bet­ter is nev­er the prototype.

    1. Aleksandr Prokudin Avatar
      Aleksandr Prokudin

      But nobody is propos­ing this work­bench for inclu­sion in FreeCAD yet

      1. barbara Avatar
        barbara

        Absolute­ly, I under­stand that! Except me, I guess: I was try­ing to pitch the idea, since I think this has the poten­tial to grow into some­thing great that ought to be includ­ed out-of-the-box 🙂

Discover more from FreeCAD News

Subscribe now to keep reading and get access to the full archive.

Continue reading