On Fri, Nov 20, 2020 at 3:49 PM Øyvind Sæbø <oyvind.s...@gmail.com> wrote:
> Yes, I think it makes sense to continue with the general discussion in the > meantime:) It could also help make sure that I don't develop Gremlint in a > direction which will impose unnecessary restrictions in the future. I've > included some of my thoughts to your questions below: > > 1. > Gremlint is currently an installable JavaScript package, so for all > JavaScript projects using npm it can already be installed and called as a > function. I imagine it would also be easy to set up a Gremlint-as-a-service > API with Node.js to work with non-JavaScript projects. > At the risk of greater complexity to documentation generation perhaps a local nodejs service might work. A shame nashorn never quite took off or else we could have perhaps embedded it for more direct usage with Gremlin Console. I wonder if we could actually get rid of Gremlin Console usage though if there was a more convenient way to do formatting/translation. By the way, can gremlint still run completely in browser or does it need nodejs at this point? > 2 & 4. > I think Gremlint as a translator could make sense. Currently it essentially > just translates Groovy-Gremlin to an (almost) language-agnostic syntax > tree, adds formatting information based on a config object and then > recreates a Groovy Gremlin query. But it could just as well recreate the > query as a JavaScript or Python Gremlin query. I think it would be > relatively easy to make Gremlint translate from any Groovy-esque Gremlin > implementation to another Gremlin implementation by simply replacing the > function which is currently used to recreate the query from the formatted > syntax tree. > Would it help make any of your work easier if there were an actual antlr grammar for Gremlin? > 3. > I've already tried to keep Gremlint a bit language agnostic by leaving all > non-Gremlin code around and between queries unaltered. In most cases, any > code Gremlint doesn't understand inside a query is just handled as if it > was a single word, and left as is. I'm not too familiar with > language-specific differences between different Gremlin-implementations, > but it is my impression that it should already handle JavaScript Gremlin > quite well. For .NET I assume I would need to add support for more than > just single-quote strings (this would come in handy for all languages), and > for Python Gremlin it would need to handle backslashes between each new > line of the query. Gremlin for Clojure would probably be more challenging > (fun fact: The Gremlint parser is actually based on a Clojure parser I > wrote earlier). > .NET might not be hard to format but I sense translation will not be easy given how easily Gremlin loses types > Food for thought: > One thing I've been a bit back and forth on is whether it would be > preferable if Gremlint is super-robust and accepts and formats any language > you pass it with minimal configuration, or if it should only accept one > Gremlin implementation "at once" and require the user to manually specify > the language of their input. > >From the perspective of a user interface, I don't think it's necessary to have too much magic. Ifi t could simply detect a language and make a good guess at it and then allow users more fine grained control from there that would be pretty impressive, but if they had to wholly make the step of selecting the language themselves I don't think that's terrible. > > fre. 20. nov. 2020 kl. 14:53 skrev Stephen Mallette <spmalle...@gmail.com > >: > > > Thanks for another update. Nice to know you're still moving along with > > improvements there. Perhaps we could continue with general discussions > > about Gremlint while we wait for legal issues to settle? For instance, > one > > of the things on my mind is how to make gremlint functionality as > available > > as possible to the various use cases we say each day. Obviously, it > > currently handles the most common one of providing a place to paste some > > Gremlin in a UI and get back formatted Gremlin. What about other things: > > > > 1. there are many Gremlin "development environments (e.g. gremlify, > Gremlin > > Console, notebooks, etc) - how can gremlint be available to them? > > 2. we are working on Gremlin translators in all languages - which might > > actually allow us to generate documentation using a single source for > > Gremlin with translation to all the others. But, that can only work if we > > have something like Gremlint to format those translations. How would > > gremlint fit in there? > > 3. gremlint (i think) is generally good at formatting the Groovy/Java > > syntax - what will it entail to make it so that it's capable of dealing > > with other languages like Python, .NET etc. > > 4. thinking more about 2, I wonder how gremlint fits with translators. > the > > two seem quite related to me, but i'm not sure what that means. it seems > > cool if you could take a bit of Gremlin Groovy, paste it into gremlint > then > > translate/lint to Gremlin Javascript. > > > > I'm just posing questions here for us to discuss a bit. Answers are great > > but more questions would also be welcome if anyone has any. > > > > > > > > On Fri, Nov 20, 2020 at 2:11 AM Øyvind Sæbø <oyvind.s...@gmail.com> > wrote: > > > > > Hello everyone and happy November. > > > > > > I requested another update on when the Gremlint paperwork process would > > be > > > done yesterday. It seems like it has been pushed back once again, the > new > > > estimate being two weeks from now. > > > > > > I figured I should give some context on why this is taking longer than > > > expected. After all, filling out the ICLA and CCLA shouldn't take more > > than > > > a day. Since I developed Gremlint in my spare time, and we want to > donate > > > it on behalf of Ardoq, where I work, we have an internal IP transfer > > > process as well, and this is what has been dragging out. > > > > > > In the meantime I've been trying to make the best of the situation by > > > working my way through the remaining Gremlint issues on GitHub. As part > > of > > > that I've rewritten Gremlint <https://github.com/OyvindSabo/gremlint> > to > > > be > > > a separate package which can be imported in the Gremlint.com website, > > which > > > now resides in a different repository > > > <https://github.com/OyvindSabo/gremlint.com>. > > > > > > Let's hope I can report on more progress next time. > > > > > >
