I think we can call it logical. Since we currently only use it to generate query operators (if we treat column transformers as operators too), we may just call the module query-operators.
It is also okay to improve the plugin, but I know very little about it and am not sure if it is possible or how many efforts it will require. Best, Tian Jiang ---- Replied Message ---- | From | Christofer Dutz<[email protected]> | | Date | 2/12/2026 03:59 | | To | [email protected]<[email protected]> | | Subject | AW: PROPOSAL: move freemarker to a separate module | Well … Drill is an Apache project … how about contacting the community and submitting a PR? Doesn’t have to be a fork. Chris Von: Haonan Hou <[email protected]> Datum: Mittwoch, 11. Februar 2026 um 16:03 An: [email protected] <[email protected]> Betreff: Re: PROPOSAL: move freemarker to a separate module Hi, Is it possible to implement a freemark plugin that supports parallel compilation by ourselves? We may fork the code of drill tool and make some improvement. Haonan On 2026/02/10 09:40:18 Tian Jiang wrote: Greetings everyone, We are using freemarker to generate operators for query processing in the DataNode module, which is the most complex module and consumes most of the time during compilation. Before I can run some ITs, especially Pipe-related ones, I must rerun maven to compile the changes. The long compiling time is very frustrating. One way is to enable parallel compilation with "-T" maven option. However, the freemarker plugin forbids parallel compilation, causing the DataNode module not to be compiled in parallel. Thus, I propose to move the freemarker and the generated code into a separate module to enable the parallel compilation of the DataNode module, saving everyone's precious time. Best, Tian Jiang
