1. 7
  1.  

  2. 2

    I did something fairly similar in perl decades ago… so I know this is Good Idea….

    ….and looks to be a much better implementation than mine. (ps: If you have very looong running queries this style plays very nice with “make” to only update things that change.)

    Do you have a handy “compile them all and install them” script somewhere…..

    … I seem to be lost in a maze of this needs that and that is not installed.

    1. 2

      I did something fairly similar in perl decades ago… so I know this is Good Idea…. ….and looks to be a much better implementation than mine.

      Thanks.

      If you have very looong running queries this style plays very nice with “make” to only update things that change.

      Yes, make is very useful. BTW: this is one of reasons why we do not invent a new shell or language etc. – and rather integrate with existing ecosystem and use good old unix way – thus it can be called not only from Bash, but also from any other shell or Make.

      Do you have a handy “compile them all and install them” script somewhere…..

      There is installation recipe in the release notes.

      Please note that it is still a development version (it has unstable API). There will be standard distribution packages when we reach version 1.0. Currently, the installation is bit uncomfortable… (but you see exactly what it does – there is no magic black box script).

      Relational pipes are very modular. Which is a way to reduce the complexity. It has just few hundreds lines of code per module and most modules depend only on the standard library and the compiler. Thus you do not have to trust us – you can audit whole source code in reasonable time, because you can choose only modules (and dependencies) you really need. If you e.g. do not need XML modules, you will not depend on XML libraries. The same is for Guile or SQLite modules etc. Relational pipes can bridge various technologies, but it does not force you to depend on them – it is optional.