Alex said: "There seem to be a lot of people starting projects to re-create OpenSCAD, in their own way, and none is anything like as compatible as I would need them to be for them to be useful. Even OpenJSCAD and FreeCAD rarely open one of my designs from OpenSCAD..." After spending a few years on the "OpenSCAD 2" project, my experience is that creating a new language implementation with 100% backwards compatibility with OpenSCAD is an extreme technical challenge, and I eventually abandoned it as a goal. The problem is that the language is so full of quirks that are not found in any other language, many of which are undocumented. For one thing, the scoping rules are so weird that you can't "transpile" OpenSCAD into a conventional language or compile into VM code for a conventional language, and preserve the semantics. In my opinion, the practical way to achieve full backwards compatibility is to use the existing OpenSCAD interpreter to compile an OpenSCAD script into a CSG tree, then write a back end that processes that CSG tree. Mixotricha's approach leads to an OpenSCAD-like language that isn't fully backwards compatible. You won't be able to take arbitrary OpenSCAD scripts off Thingiverse or whatever and expect them to just work without being "ported" to MakerTron. However, if you are writing new code from scratch, then you should be able to write code that is portable between OpenSCAD and MakerTron. The first example I tried was: cylinder(r=1,h=5,$fs=6); This is the OpenSCAD idiom for a hexagonal prism. What I got instead was a cylinder with a circular cross section. I also tried x = 1; function x(a) = a + 1; module x(a) cube(x(x)); x(2); but I didn't get a result back, and lost contact with the makertron server. On 19 July 2017 at 05:13, Alex Gibson <alex@alexgibson.net> wrote: > Hi there Mixotricha, > > I was quite excited but I tried to copy-paste 3 Openscad files of sharply > decreasing complexity, and none of them would produce an object, or even an > error. > > When I refreshed the page, I couldn't get back to the original example > file that first greeted me - so maybe a browser caching issue? > > There seem to be a lot of people starting projects to re-create OpenSCAD, > in their own way, and none is anything like as compatible as I would need > them to be for them to be useful. Even OpenJSCAD and FreeCAD rarely open > one of my designs from OpenSCAD... > > For example, supporting 'minkowski' is an absolute must, intersections, > linear extrudes, logic and ideally text functions - some way to handle > fonts could complicate this but it would be worth it! How would we call a > function from a separate file? That sort of thing... > > Cheers > Alex > > -----Original Message----- > From: Discuss [mailto:discuss-bounces@lists.openscad.org] On Behalf Of > MichaelAtOz > Sent: 19 July 2017 03:27 > To: discuss@lists.openscad.org > Subject: Re: [OpenSCAD] Project Announcement Makertron > > mixotricha wrote > > Hello All, > > > > I'd like to announce Makertron. > > http://makertron.io <http://makertron.io> > > > > It is an alternative I am working on that addresses some of problems I > > see with OpenSCAD and with OpenJSCAD. > > > > It is a bit like ImplicitCad in that it has a server back end -but- it > > also has a client front end. > > > > I have seen a few people have a go at this but none of them were quite > > doing what I needed or had some core implementation problems. > > > > I've gone after some of those and tried to resolve them. > > > > I'd love to get some feedback and would be happy to help answer any > > questions about my project. > > > > Greetings From Australia > > > > Damien > > "At least 90% of the OpenSCAD language specification is now supported" > > Is that documented? > > > > > ----- > Admin - PM me if you need anything, or if I've done something stupid... > > Unless specifically shown otherwise above, my contribution is in the > Public Domain; to the extent possible under law, I have waived all > copyright and related or neighbouring rights to this work. Obviously > inclusion of works of previous authors is not included in the above. > > The TPP is no simple “trade agreement.” Fight it! > http://www.ourfairdeal.org/ time is running out! > -- > View this message in context: http://forum.openscad.org/ > Project-Announcement-Makertron-tp21849p21853.html > Sent from the OpenSCAD mailing list archive at Nabble.com. > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

Sorry, I accidently sent that message before I was finished writing it, and it contained a typo, i wrote $fs where i meant $fn. On 19 July 2017 at 13:18, doug moen <doug@moens.org> wrote: > Alex said: "There seem to be a lot of people starting projects to > re-create OpenSCAD, in their own way, and none is anything like as > compatible as I would need them to be for them to be useful. Even > OpenJSCAD and FreeCAD rarely open one of my designs from OpenSCAD..." > > After spending a few years on the "OpenSCAD 2" project, my experience is > that creating a new language implementation with 100% backwards > compatibility with OpenSCAD is an extreme technical challenge, and I > eventually abandoned it as a goal. The problem is that the language is so > full of quirks that are not found in any other language, many of which are > undocumented. For one thing, the scoping rules are so weird that you can't > "transpile" OpenSCAD into a conventional language or compile into VM code > for a conventional language, and preserve the semantics. > > In my opinion, the practical way to achieve full backwards compatibility > is to use the existing OpenSCAD interpreter to compile an OpenSCAD script > into a CSG tree, then write a back end that processes that CSG tree. > > Mixotricha's approach leads to an OpenSCAD-like language that isn't fully > backwards compatible. You won't be able to take arbitrary OpenSCAD scripts > off Thingiverse or whatever and expect them to just work without being > "ported" to MakerTron. However, if you are writing new code from scratch, > then you should be able to write code that is portable between OpenSCAD and > MakerTron. > > The first example I tried was: > cylinder(r=1,h=5,$fs=6); > This is the OpenSCAD idiom for a hexagonal prism. What I got instead was a > cylinder with a circular cross section. > > I also tried > x = 1; > function x(a) = a + 1; > module x(a) cube(x(x)); > x(2); > but I didn't get a result back, and lost contact with the makertron server. > > > On 19 July 2017 at 05:13, Alex Gibson <alex@alexgibson.net> wrote: > >> Hi there Mixotricha, >> >> I was quite excited but I tried to copy-paste 3 Openscad files of sharply >> decreasing complexity, and none of them would produce an object, or even an >> error. >> >> When I refreshed the page, I couldn't get back to the original example >> file that first greeted me - so maybe a browser caching issue? >> >> There seem to be a lot of people starting projects to re-create OpenSCAD, >> in their own way, and none is anything like as compatible as I would need >> them to be for them to be useful. Even OpenJSCAD and FreeCAD rarely open >> one of my designs from OpenSCAD... >> >> For example, supporting 'minkowski' is an absolute must, intersections, >> linear extrudes, logic and ideally text functions - some way to handle >> fonts could complicate this but it would be worth it! How would we call a >> function from a separate file? That sort of thing... >> >> Cheers >> Alex >> >> -----Original Message----- >> From: Discuss [mailto:discuss-bounces@lists.openscad.org] On Behalf Of >> MichaelAtOz >> Sent: 19 July 2017 03:27 >> To: discuss@lists.openscad.org >> Subject: Re: [OpenSCAD] Project Announcement Makertron >> >> mixotricha wrote >> > Hello All, >> > >> > I'd like to announce Makertron. >> > http://makertron.io <http://makertron.io> >> > >> > It is an alternative I am working on that addresses some of problems I >> > see with OpenSCAD and with OpenJSCAD. >> > >> > It is a bit like ImplicitCad in that it has a server back end -but- it >> > also has a client front end. >> > >> > I have seen a few people have a go at this but none of them were quite >> > doing what I needed or had some core implementation problems. >> > >> > I've gone after some of those and tried to resolve them. >> > >> > I'd love to get some feedback and would be happy to help answer any >> > questions about my project. >> > >> > Greetings From Australia >> > >> > Damien >> >> "At least 90% of the OpenSCAD language specification is now supported" >> >> Is that documented? >> >> >> >> >> ----- >> Admin - PM me if you need anything, or if I've done something stupid... >> >> Unless specifically shown otherwise above, my contribution is in the >> Public Domain; to the extent possible under law, I have waived all >> copyright and related or neighbouring rights to this work. Obviously >> inclusion of works of previous authors is not included in the above. >> >> The TPP is no simple “trade agreement.” Fight it! >> http://www.ourfairdeal.org/ time is running out! >> -- >> View this message in context: http://forum.openscad.org/Proj >> ect-Announcement-Makertron-tp21849p21853.html >> Sent from the OpenSCAD mailing list archive at Nabble.com. >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> > >

In my opinion, the practical way to achieve full backwards compatibility is to use the existing OpenSCAD interpreter to compile an OpenSCAD script into a CSG tree, then write a back end that processes that CSG tree. If you are interested my previous work on this did exactly that. It is in my humbaba tree. I went through this. Putting a socket layer on OpenSCAD is trivial. Building it in to a shared library so I could access it through nodeJS also trivial. But then what you will find is issues with speed and the complex pathway for multiple rendering targets. The roadblock that emerged was interaction between a client side interface in something like react and dealing with the OpenCSG/Goldfeather approach. Making OpenSCAD headless is not hard. But the rest of it will rapidly lead you to a series of painful roadblocks. I've got a better way. Think transpiler. Not parser. The examples you have given above are excellent and also things that can be fixed. I've had the client up live for one day and this sort of input is exactly what I needed. Keep them coming. On Thu, Jul 20, 2017 at 3:25 AM, doug moen <doug@moens.org> wrote: > Sorry, I accidently sent that message before I was finished writing it, > and it contained a typo, i wrote $fs where i meant $fn. > > On 19 July 2017 at 13:18, doug moen <doug@moens.org> wrote: > >> Alex said: "There seem to be a lot of people starting projects to >> re-create OpenSCAD, in their own way, and none is anything like as >> compatible as I would need them to be for them to be useful. Even >> OpenJSCAD and FreeCAD rarely open one of my designs from OpenSCAD..." >> >> After spending a few years on the "OpenSCAD 2" project, my experience is >> that creating a new language implementation with 100% backwards >> compatibility with OpenSCAD is an extreme technical challenge, and I >> eventually abandoned it as a goal. The problem is that the language is so >> full of quirks that are not found in any other language, many of which are >> undocumented. For one thing, the scoping rules are so weird that you can't >> "transpile" OpenSCAD into a conventional language or compile into VM code >> for a conventional language, and preserve the semantics. >> >> In my opinion, the practical way to achieve full backwards compatibility >> is to use the existing OpenSCAD interpreter to compile an OpenSCAD script >> into a CSG tree, then write a back end that processes that CSG tree. >> >> Mixotricha's approach leads to an OpenSCAD-like language that isn't fully >> backwards compatible. You won't be able to take arbitrary OpenSCAD scripts >> off Thingiverse or whatever and expect them to just work without being >> "ported" to MakerTron. However, if you are writing new code from scratch, >> then you should be able to write code that is portable between OpenSCAD and >> MakerTron. >> >> The first example I tried was: >> cylinder(r=1,h=5,$fs=6); >> This is the OpenSCAD idiom for a hexagonal prism. What I got instead was >> a cylinder with a circular cross section. >> >> I also tried >> x = 1; >> function x(a) = a + 1; >> module x(a) cube(x(x)); >> x(2); >> but I didn't get a result back, and lost contact with the makertron >> server. >> >> >> On 19 July 2017 at 05:13, Alex Gibson <alex@alexgibson.net> wrote: >> >>> Hi there Mixotricha, >>> >>> I was quite excited but I tried to copy-paste 3 Openscad files of >>> sharply decreasing complexity, and none of them would produce an object, or >>> even an error. >>> >>> When I refreshed the page, I couldn't get back to the original example >>> file that first greeted me - so maybe a browser caching issue? >>> >>> There seem to be a lot of people starting projects to re-create >>> OpenSCAD, in their own way, and none is anything like as compatible as I >>> would need them to be for them to be useful. Even OpenJSCAD and FreeCAD >>> rarely open one of my designs from OpenSCAD... >>> >>> For example, supporting 'minkowski' is an absolute must, intersections, >>> linear extrudes, logic and ideally text functions - some way to handle >>> fonts could complicate this but it would be worth it! How would we call a >>> function from a separate file? That sort of thing... >>> >>> Cheers >>> Alex >>> >>> -----Original Message----- >>> From: Discuss [mailto:discuss-bounces@lists.openscad.org] On Behalf Of >>> MichaelAtOz >>> Sent: 19 July 2017 03:27 >>> To: discuss@lists.openscad.org >>> Subject: Re: [OpenSCAD] Project Announcement Makertron >>> >>> mixotricha wrote >>> > Hello All, >>> > >>> > I'd like to announce Makertron. >>> > http://makertron.io <http://makertron.io> >>> > >>> > It is an alternative I am working on that addresses some of problems I >>> > see with OpenSCAD and with OpenJSCAD. >>> > >>> > It is a bit like ImplicitCad in that it has a server back end -but- it >>> > also has a client front end. >>> > >>> > I have seen a few people have a go at this but none of them were quite >>> > doing what I needed or had some core implementation problems. >>> > >>> > I've gone after some of those and tried to resolve them. >>> > >>> > I'd love to get some feedback and would be happy to help answer any >>> > questions about my project. >>> > >>> > Greetings From Australia >>> > >>> > Damien >>> >>> "At least 90% of the OpenSCAD language specification is now supported" >>> >>> Is that documented? >>> >>> >>> >>> >>> ----- >>> Admin - PM me if you need anything, or if I've done something stupid... >>> >>> Unless specifically shown otherwise above, my contribution is in the >>> Public Domain; to the extent possible under law, I have waived all >>> copyright and related or neighbouring rights to this work. Obviously >>> inclusion of works of previous authors is not included in the above. >>> >>> The TPP is no simple “trade agreement.” Fight it! >>> http://www.ourfairdeal.org/ time is running out! >>> -- >>> View this message in context: http://forum.openscad.org/Proj >>> ect-Announcement-Makertron-tp21849p21853.html >>> Sent from the OpenSCAD mailing list archive at Nabble.com. >>> >>> _______________________________________________ >>> OpenSCAD mailing list >>> Discuss@lists.openscad.org >>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>> >>> >>> _______________________________________________ >>> OpenSCAD mailing list >>> Discuss@lists.openscad.org >>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>> >> >> > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > > -- Damien Towning CEO / Developer http://makertron.io

I also forgot to say also that having made the OpenSCAD headless it brings one to thinking about moving the parser as a whole. After all the grammar is already provided. But as I am sure you know ( and as is often the case with any grammar ) moving the OpenSCAD grammar as it currently stands to a new target is not at all trivial. So I changed my thinking. Don't write a parser. Write a transpiler. Everything OpenSCAD does results in a stack of geometrical operations. You can get to those operations via a transpile. You can also address quirks in a really powerful way in a transpile. The example I would give of that is the lazy OpenSCAD closure. On Thu, Jul 20, 2017 at 10:58 AM, Damien Towning <connolly.damien@gmail.com> wrote: > In my opinion, the practical way to achieve full backwards compatibility > is to use the existing OpenSCAD interpreter to compile an OpenSCAD script > into a CSG tree, then write a back end that processes that CSG tree. > > If you are interested my previous work on this did exactly that. It is in > my humbaba tree. I went through this. Putting a socket layer on OpenSCAD is > trivial. Building it in to a shared library so I could access it through > nodeJS also trivial. But then what you will find is issues with speed and > the complex pathway for multiple rendering targets. The roadblock that > emerged was interaction between a client side interface in something like > react and dealing with the OpenCSG/Goldfeather approach. Making OpenSCAD > headless is not hard. But the rest of it will rapidly lead you to a series > of painful roadblocks. I've got a better way. Think transpiler. Not parser. > The examples you have given above are excellent and also things that can be > fixed. I've had the client up live for one day and this sort of input is > exactly what I needed. Keep them coming. > > > > > On Thu, Jul 20, 2017 at 3:25 AM, doug moen <doug@moens.org> wrote: > >> Sorry, I accidently sent that message before I was finished writing it, >> and it contained a typo, i wrote $fs where i meant $fn. >> >> On 19 July 2017 at 13:18, doug moen <doug@moens.org> wrote: >> >>> Alex said: "There seem to be a lot of people starting projects to >>> re-create OpenSCAD, in their own way, and none is anything like as >>> compatible as I would need them to be for them to be useful. Even >>> OpenJSCAD and FreeCAD rarely open one of my designs from OpenSCAD..." >>> >>> After spending a few years on the "OpenSCAD 2" project, my experience is >>> that creating a new language implementation with 100% backwards >>> compatibility with OpenSCAD is an extreme technical challenge, and I >>> eventually abandoned it as a goal. The problem is that the language is so >>> full of quirks that are not found in any other language, many of which are >>> undocumented. For one thing, the scoping rules are so weird that you can't >>> "transpile" OpenSCAD into a conventional language or compile into VM code >>> for a conventional language, and preserve the semantics. >>> >>> In my opinion, the practical way to achieve full backwards compatibility >>> is to use the existing OpenSCAD interpreter to compile an OpenSCAD script >>> into a CSG tree, then write a back end that processes that CSG tree. >>> >>> Mixotricha's approach leads to an OpenSCAD-like language that isn't >>> fully backwards compatible. You won't be able to take arbitrary OpenSCAD >>> scripts off Thingiverse or whatever and expect them to just work without >>> being "ported" to MakerTron. However, if you are writing new code from >>> scratch, then you should be able to write code that is portable between >>> OpenSCAD and MakerTron. >>> >>> The first example I tried was: >>> cylinder(r=1,h=5,$fs=6); >>> This is the OpenSCAD idiom for a hexagonal prism. What I got instead was >>> a cylinder with a circular cross section. >>> >>> I also tried >>> x = 1; >>> function x(a) = a + 1; >>> module x(a) cube(x(x)); >>> x(2); >>> but I didn't get a result back, and lost contact with the makertron >>> server. >>> >>> >>> On 19 July 2017 at 05:13, Alex Gibson <alex@alexgibson.net> wrote: >>> >>>> Hi there Mixotricha, >>>> >>>> I was quite excited but I tried to copy-paste 3 Openscad files of >>>> sharply decreasing complexity, and none of them would produce an object, or >>>> even an error. >>>> >>>> When I refreshed the page, I couldn't get back to the original example >>>> file that first greeted me - so maybe a browser caching issue? >>>> >>>> There seem to be a lot of people starting projects to re-create >>>> OpenSCAD, in their own way, and none is anything like as compatible as I >>>> would need them to be for them to be useful. Even OpenJSCAD and FreeCAD >>>> rarely open one of my designs from OpenSCAD... >>>> >>>> For example, supporting 'minkowski' is an absolute must, intersections, >>>> linear extrudes, logic and ideally text functions - some way to handle >>>> fonts could complicate this but it would be worth it! How would we call a >>>> function from a separate file? That sort of thing... >>>> >>>> Cheers >>>> Alex >>>> >>>> -----Original Message----- >>>> From: Discuss [mailto:discuss-bounces@lists.openscad.org] On Behalf Of >>>> MichaelAtOz >>>> Sent: 19 July 2017 03:27 >>>> To: discuss@lists.openscad.org >>>> Subject: Re: [OpenSCAD] Project Announcement Makertron >>>> >>>> mixotricha wrote >>>> > Hello All, >>>> > >>>> > I'd like to announce Makertron. >>>> > http://makertron.io <http://makertron.io> >>>> > >>>> > It is an alternative I am working on that addresses some of problems I >>>> > see with OpenSCAD and with OpenJSCAD. >>>> > >>>> > It is a bit like ImplicitCad in that it has a server back end -but- it >>>> > also has a client front end. >>>> > >>>> > I have seen a few people have a go at this but none of them were quite >>>> > doing what I needed or had some core implementation problems. >>>> > >>>> > I've gone after some of those and tried to resolve them. >>>> > >>>> > I'd love to get some feedback and would be happy to help answer any >>>> > questions about my project. >>>> > >>>> > Greetings From Australia >>>> > >>>> > Damien >>>> >>>> "At least 90% of the OpenSCAD language specification is now supported" >>>> >>>> Is that documented? >>>> >>>> >>>> >>>> >>>> ----- >>>> Admin - PM me if you need anything, or if I've done something stupid... >>>> >>>> Unless specifically shown otherwise above, my contribution is in the >>>> Public Domain; to the extent possible under law, I have waived all >>>> copyright and related or neighbouring rights to this work. Obviously >>>> inclusion of works of previous authors is not included in the above. >>>> >>>> The TPP is no simple “trade agreement.” Fight it! >>>> http://www.ourfairdeal.org/ time is running out! >>>> -- >>>> View this message in context: http://forum.openscad.org/Proj >>>> ect-Announcement-Makertron-tp21849p21853.html >>>> Sent from the OpenSCAD mailing list archive at Nabble.com. >>>> >>>> _______________________________________________ >>>> OpenSCAD mailing list >>>> Discuss@lists.openscad.org >>>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>>> >>>> >>>> _______________________________________________ >>>> OpenSCAD mailing list >>>> Discuss@lists.openscad.org >>>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>>> >>> >>> >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> >> > > > -- > Damien Towning > CEO / Developer > http://makertron.io > > -- Damien Towning CEO / Developer http://makertron.io

I'm having difficulty following you; we may be using terminology differently. OpenSCAD proceses a script using the following steps: * the parser converts the script to a parse tree * the evaluator converts the parse tree to a CSG tree * the renderer converts the CSG tree to the final product. OpenSCAD has a previewer that uses OpenCSG, and a mesh generator that uses CGAL. You can also dump the CSG tree as text (a simplified OpenSCAD script), without further processing. A transpiler is basically a compiler that converts source code in one language to source code in another language. A transpiler will always have a parser as the first step of processing. So I don't know what you mean by "Don't write a parser. Write a transpiler." In OpenSCAD, the parser is not very difficult to write. But it is extremely difficult to transpile OpenSCAD into another language without first using the OpenSCAD evaluator to construct a CSG tree. That's because of the very unusual dynamic scoping rules used by the evaluator. Within a function or module, the meaning of an identifier is only known at run time. A transpiler would either interpret identifiers using static scoping rules, which doesn't work, or it would have to transpile into another language that uses the exact same dynamic scoping rules (which aren't documented, and which I never fully succeeded in reverse engineering.) What you want to do is transpile the CSG tree into your final format. The CSG tree also has OpenSCAD syntax, but it's a subset that eliminates all of the hard problems, so you can feed that into your existing transpiler. On 19 July 2017 at 20:58, Damien Towning <connolly.damien@gmail.com> wrote: > In my opinion, the practical way to achieve full backwards compatibility > is to use the existing OpenSCAD interpreter to compile an OpenSCAD script > into a CSG tree, then write a back end that processes that CSG tree. > > If you are interested my previous work on this did exactly that. It is in > my humbaba tree. I went through this. Putting a socket layer on OpenSCAD is > trivial. Building it in to a shared library so I could access it through > nodeJS also trivial. But then what you will find is issues with speed and > the complex pathway for multiple rendering targets. The roadblock that > emerged was interaction between a client side interface in something like > react and dealing with the OpenCSG/Goldfeather approach. Making OpenSCAD > headless is not hard. But the rest of it will rapidly lead you to a series > of painful roadblocks. I've got a better way. Think transpiler. Not parser. > The examples you have given above are excellent and also things that can be > fixed. I've had the client up live for one day and this sort of input is > exactly what I needed. Keep them coming. > > > > > On Thu, Jul 20, 2017 at 3:25 AM, doug moen <doug@moens.org> wrote: > >> Sorry, I accidently sent that message before I was finished writing it, >> and it contained a typo, i wrote $fs where i meant $fn. >> >> On 19 July 2017 at 13:18, doug moen <doug@moens.org> wrote: >> >>> Alex said: "There seem to be a lot of people starting projects to >>> re-create OpenSCAD, in their own way, and none is anything like as >>> compatible as I would need them to be for them to be useful. Even >>> OpenJSCAD and FreeCAD rarely open one of my designs from OpenSCAD..." >>> >>> After spending a few years on the "OpenSCAD 2" project, my experience is >>> that creating a new language implementation with 100% backwards >>> compatibility with OpenSCAD is an extreme technical challenge, and I >>> eventually abandoned it as a goal. The problem is that the language is so >>> full of quirks that are not found in any other language, many of which are >>> undocumented. For one thing, the scoping rules are so weird that you can't >>> "transpile" OpenSCAD into a conventional language or compile into VM code >>> for a conventional language, and preserve the semantics. >>> >>> In my opinion, the practical way to achieve full backwards compatibility >>> is to use the existing OpenSCAD interpreter to compile an OpenSCAD script >>> into a CSG tree, then write a back end that processes that CSG tree. >>> >>> Mixotricha's approach leads to an OpenSCAD-like language that isn't >>> fully backwards compatible. You won't be able to take arbitrary OpenSCAD >>> scripts off Thingiverse or whatever and expect them to just work without >>> being "ported" to MakerTron. However, if you are writing new code from >>> scratch, then you should be able to write code that is portable between >>> OpenSCAD and MakerTron. >>> >>> The first example I tried was: >>> cylinder(r=1,h=5,$fs=6); >>> This is the OpenSCAD idiom for a hexagonal prism. What I got instead was >>> a cylinder with a circular cross section. >>> >>> I also tried >>> x = 1; >>> function x(a) = a + 1; >>> module x(a) cube(x(x)); >>> x(2); >>> but I didn't get a result back, and lost contact with the makertron >>> server. >>> >>> >>> On 19 July 2017 at 05:13, Alex Gibson <alex@alexgibson.net> wrote: >>> >>>> Hi there Mixotricha, >>>> >>>> I was quite excited but I tried to copy-paste 3 Openscad files of >>>> sharply decreasing complexity, and none of them would produce an object, or >>>> even an error. >>>> >>>> When I refreshed the page, I couldn't get back to the original example >>>> file that first greeted me - so maybe a browser caching issue? >>>> >>>> There seem to be a lot of people starting projects to re-create >>>> OpenSCAD, in their own way, and none is anything like as compatible as I >>>> would need them to be for them to be useful. Even OpenJSCAD and FreeCAD >>>> rarely open one of my designs from OpenSCAD... >>>> >>>> For example, supporting 'minkowski' is an absolute must, intersections, >>>> linear extrudes, logic and ideally text functions - some way to handle >>>> fonts could complicate this but it would be worth it! How would we call a >>>> function from a separate file? That sort of thing... >>>> >>>> Cheers >>>> Alex >>>> >>>> -----Original Message----- >>>> From: Discuss [mailto:discuss-bounces@lists.openscad.org] On Behalf Of >>>> MichaelAtOz >>>> Sent: 19 July 2017 03:27 >>>> To: discuss@lists.openscad.org >>>> Subject: Re: [OpenSCAD] Project Announcement Makertron >>>> >>>> mixotricha wrote >>>> > Hello All, >>>> > >>>> > I'd like to announce Makertron. >>>> > http://makertron.io <http://makertron.io> >>>> > >>>> > It is an alternative I am working on that addresses some of problems I >>>> > see with OpenSCAD and with OpenJSCAD. >>>> > >>>> > It is a bit like ImplicitCad in that it has a server back end -but- it >>>> > also has a client front end. >>>> > >>>> > I have seen a few people have a go at this but none of them were quite >>>> > doing what I needed or had some core implementation problems. >>>> > >>>> > I've gone after some of those and tried to resolve them. >>>> > >>>> > I'd love to get some feedback and would be happy to help answer any >>>> > questions about my project. >>>> > >>>> > Greetings From Australia >>>> > >>>> > Damien >>>> >>>> "At least 90% of the OpenSCAD language specification is now supported" >>>> >>>> Is that documented? >>>> >>>> >>>> >>>> >>>> ----- >>>> Admin - PM me if you need anything, or if I've done something stupid... >>>> >>>> Unless specifically shown otherwise above, my contribution is in the >>>> Public Domain; to the extent possible under law, I have waived all >>>> copyright and related or neighbouring rights to this work. Obviously >>>> inclusion of works of previous authors is not included in the above. >>>> >>>> The TPP is no simple “trade agreement.” Fight it! >>>> http://www.ourfairdeal.org/ time is running out! >>>> -- >>>> View this message in context: http://forum.openscad.org/Proj >>>> ect-Announcement-Makertron-tp21849p21853.html >>>> Sent from the OpenSCAD mailing list archive at Nabble.com. >>>> >>>> _______________________________________________ >>>> OpenSCAD mailing list >>>> Discuss@lists.openscad.org >>>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>>> >>>> >>>> _______________________________________________ >>>> OpenSCAD mailing list >>>> Discuss@lists.openscad.org >>>> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >>>> >>> >>> >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> >> > > > -- > Damien Towning > CEO / Developer > http://makertron.io > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > >

On github it says: > Lastly the OpenSCAD parser while supporting an excellent CSG language > ( though in some areas ambiguous ) is so glued to the interface and > supporting code that it is hard for the language to evolve or move > forwards from that interface. It is our opinion that this effort has > stalled and something new needed to be developed to address these > issues. > If that's one of the core premises, I'd suggest to have a look at http://apidesign.org/ Also it would be interesting to have a look at that geometry engine (quote "The Makertron server supports its own internal geometry engine based on a BREP representation.") which seems to be based on Open Cascade. I can only find shared libs in the server repo. ciao, Torsten.

Hi Alex, I pushed a new update out today with some improved argument parsing. I see that nice little example you gave now loads. I'll be sorting out some sort of supported features matrix and a time line as well. Just a lot to get done :) On Thu, Jul 20, 2017 at 4:26 PM, Torsten Paul <Torsten.Paul@gmx.de> wrote: > On github it says: > > > Lastly the OpenSCAD parser while supporting an excellent CSG language > > ( though in some areas ambiguous ) is so glued to the interface and > > supporting code that it is hard for the language to evolve or move > > forwards from that interface. It is our opinion that this effort has > > stalled and something new needed to be developed to address these > > issues. > > > If that's one of the core premises, I'd suggest to have a look at > http://apidesign.org/ > > Also it would be interesting to have a look at that geometry engine > (quote "The Makertron server supports its own internal geometry > engine based on a BREP representation.") which seems to be based on > Open Cascade. I can only find shared libs in the server repo. > > ciao, > Torsten. > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > -- Damien Towning CEO / Developer http://makertron.io

On 2017-07-21 09:24, Damien Towning wrote: > Hi Alex, > > I pushed a new update out today with some improved argument parsing. Hi, I tried the default example from behind a corporate network yesterday and today. The page opens and I can see the axes and xy-grid. I can rotate it with the left mouse button. With the right mouse button everything just disappears to the bottom right, no way to get it back. When pressing the Generate button, nothing happens. No model displayed. Same thing with Firefox and Microsoft Edge under Win10. What are the requirements to make it work? Carsten Arnholm

Hi Carsten, I've tested it in Firefox and Chrome. It works fine in those but I have not checked it in Edge. I would suggest that if it is getting the grid and axis the trouble is perhaps with talking to the server. The client opens a port to the server ( using socket.io ) on port 3000. You can test the server by going to : http://makertron.io:3000 If it is up and responding it should give you a version number back. "Makertron server version 5.0" If not take a look in your console/debug window and if you see any interesting error messages let me know. On Fri, Jul 21, 2017 at 5:52 PM, <arnholm@arnholm.org> wrote: > On 2017-07-21 09:24, Damien Towning wrote: > >> Hi Alex, >> >> I pushed a new update out today with some improved argument parsing. >> > > Hi, > > I tried the default example from behind a corporate network yesterday and > today. The page opens and I can see the axes and xy-grid. I can rotate it > with the left mouse button. With the right mouse button everything just > disappears to the bottom right, no way to get it back. When pressing the > Generate button, nothing happens. No model displayed. Same thing with > Firefox and Microsoft Edge under Win10. > > What are the requirements to make it work? > > Carsten Arnholm > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > -- Damien Towning CEO / Developer http://makertron.io

Oh also to get the default model to load you need to click on the generate button :) On Fri, Jul 21, 2017 at 6:07 PM, Damien Towning <connolly.damien@gmail.com> wrote: > Hi Carsten, > > I've tested it in Firefox and Chrome. It works fine in those but I have > not checked it in Edge. > > I would suggest that if it is getting the grid and axis the trouble is > perhaps with talking to the server. > > The client opens a port to the server ( using socket.io ) on port 3000. > > You can test the server by going to : http://makertron.io:3000 > > If it is up and responding it should give you a version number back. > > "Makertron server version 5.0" > > If not take a look in your console/debug window and if you see any > interesting error messages let me know. > > > On Fri, Jul 21, 2017 at 5:52 PM, <arnholm@arnholm.org> wrote: > >> On 2017-07-21 09:24, Damien Towning wrote: >> >>> Hi Alex, >>> >>> I pushed a new update out today with some improved argument parsing. >>> >> >> Hi, >> >> I tried the default example from behind a corporate network yesterday and >> today. The page opens and I can see the axes and xy-grid. I can rotate it >> with the left mouse button. With the right mouse button everything just >> disappears to the bottom right, no way to get it back. When pressing the >> Generate button, nothing happens. No model displayed. Same thing with >> Firefox and Microsoft Edge under Win10. >> >> What are the requirements to make it work? >> >> Carsten Arnholm >> >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> > > > > -- > Damien Towning > CEO / Developer > http://makertron.io > > -- Damien Towning CEO / Developer http://makertron.io