Sure thing. Thank you for the reminder, Alex! Toshiya
On Wed, Feb 28, 2024 at 10:04 PM Alex Porcelli <[email protected]> wrote: > Thank you very much for the detailed information, Mario and Toshiya. > > Let's use this as an opportunity to remind ourselves about the > importance of continued communication following the guidelines we > collectively defined [1]. > > [1] - > https://cwiki.apache.org/confluence/display/KIE/Communication+within+Apache > > On Tue, Feb 27, 2024 at 6:53 AM Mario Fusco <[email protected]> wrote: > > > > Hi Alex, > > > > Just to clarify the purposes of this effort and from where this > originate, consider that the development of this new ANTLR v4 based parser > for the DRL started even before the migration to Apache and its main goal > was to support the implementation of an LSP server that could allow to > develop an IDE plugin for the DRL language providing code completion and > syntax coloring. > > > > More in general however the current Drools parser is a very legacy part > of our codebase. It is implemented with a weird mix of an ANTLR v3 > generated parser with tons of handwritten code covering the most complex > edge cases of the language. After so many years it has become very hard to > maintain and evolve to the point that now I consider it one of the most > relevant contributor to Drools' technical debt. > > > > As for the risks related with the replacement of the old parser with > this new one, I honestly don't have any concern, also considering that: > > > > 1. This is a long ongoing work that we are developing on a feature > branch. It will still require a few months of development (to be > optimistic), so for sure it won't be part of the 10.0.0 release. > > > > 2. As Toshiya already mentioned, this parser is confined in its own > module and its interactions with the rest of the project are well defined > and under control. > > > > 3. The Drools test base is huge, with tens of thousands of test cases > covering each and every details and edge cases of the DRL. I'm 100% sure > that if the new parser won't be implementing correctly even only one of the > DRL specification, there will be at least one test failure evidencing the > problem and of course we will do the replacement only when the entire > Drools test suite will be green also using the new parser. > > > > I hope this clarifies the situation. > > Thanks, > > Mario > > > > On 2024/02/26 13:57:11 Alex Porcelli wrote: > > > Toshiya, > > > > > > Thank you for the heads up, however I'm a bit concerned how far this > > > implementation has gone without a previous proposal to be discussed as > > > stated in our community communication guidelines [1]. > > > > > > Although I really appreciate the effort to keep up to date with > > > technology advances (ANTLR v4 has been around for soooo many years > > > now), but I'm also concern that this will introduce complexity and > > > risk while we still improving the codebase stability (ie. issues [2] > > > [3] [4] [5]). > > > > > > So I would like to better understand what's the timeline expected to > > > see this incorporated into the codebase, and besides the library > > > upgrade, what are the goals of this change. > > > > > > Note: I'm not sure what you mean by Drools team, as far as I know we > > > are all part of a single community under KIE. Although we might have > > > more areas of specialization, I don't think we are split into > > > different teams. > > > > > > [1] - > https://cwiki.apache.org/confluence/display/KIE/Communication+within+Apache > > > [2] - https://github.com/apache/incubator-kie-issues/issues/941 > > > [3] - https://github.com/apache/incubator-kie-issues/issues/903 > > > [4] - https://github.com/apache/incubator-kie-issues/issues/968 > > > [5] - https://github.com/apache/incubator-kie-issues/issues/969 > > > > > > > > > On Mon, Feb 26, 2024 at 3:37 AM Toshiya Kobayashi > > > <[email protected]> wrote: > > > > > > > > Hello, > > > > > > > > In the drools team, we are developing a new Antlr4 based DRL parser. > The > > > > current parser is Antlr3 based and contains lots of hard-coded logic > in a > > > > generated java code (DRL6Parser.java), so it's hard to maintain. > > > > > > > > Originally the new parser was developed under > > > > https://github.com/kiegroup/drools-lsp/ at the early stage. Now we > are > > > > testing the new DRL parser with the existing drool unit tests using a > > > > feature branch `dev-new-parser`. We hit lots of test failures (as > expected) > > > > and are fixing them one-by-one. > > > > > > > > This is the parent issue where Jiri Locker made a great effort to > split > > > > child issues (test failures): > > > > > > > > https://github.com/apache/incubator-kie-drools/issues/5678 > > > > > > > > Those child issues are fine grained, so it would be nice to > contribute! > > > > > > > > If you are willing to contribute, please put you as an assignee (if > you > > > > have a committer permission) or post a comment to avoid duplicate > work on > > > > the child issue. > > > > > > > > Also you can track the progress by watching the parent issue and > share your > > > > thoughts any time you want. > > > > > > > > Cheers, > > > > Toshiya > > > > > > --------------------------------------------------------------------- > > > To unsubscribe, e-mail: [email protected] > > > For additional commands, e-mail: [email protected] > > > > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
