Hague Grant Application: PL/Perl6 Infrastructure and Improvements
Mon, 21-Nov-2011 by
Karen Pauley
edit post
<![CDATA[
We have received the following "Hague Grant":http://www.perlfoundation.org/ian_hague_perl_6_development_grants application from Jonathan "Duke" Leto.
Before we vote on this proposal we would like to have a period of community consultation for 10 days. Please leave feedback in the comments or if you prefer send email with your comments to karen at perlfoundation.org.
**Name**
Jonathan "Duke" Leto
**Project Title**
PL/Perl6 Infrastructure and Improvements
**Synopsis**
The proposed grant would support the development of PL/Perl6, which allows people to use Perl 6 as a stored procedure language in the PostgreSQL database.
**Benefits to Perl 6 Development**
It will allow a very new and large niche of developers access to Perl 6 and provide a use case for it.
**Deliverables**
1) PL/Perl6 working on the latest stable release of Rakudo
There is currently a bug which causes some PL/Perl6 tests to crash Postgres. This could be from something that changed in Parrot, Rakudo or both. Because there was no system testing PL/Perl6 on newer releases of Rakudo, this got broken and has yet to be fixed.
2) A continuous integration setup such that PL/Perl6 is tested against newer versions of Parrot + Rakudo, so that 1) does not happen again.
This will probably be implemented on the GCC Compile Farm and/or the OSUOSL Supercell project. Documentation about how to install and maintain this CI environment will be included.
3) User-friendly web documentation for developers wanting to use PL/Perl6
This will include one set of documentation that is an overview and intro the API, and another set of documentation which gives various examples, such as using a Perl 6 Grammar in a stored procedure.
4) User-friendly web documentation for installing/compiling PL/Perl6
This documentation will guide people on how to compile PL/Perl6 from a release tarball, from the Git repo or from a package manager.
5) Screencast showing how to write a stored procedure in Perl 6
A screencast that can be used for promoting Perl 6 in PostgreSQL will also be a deliverable. This will show how to write a stored procedure in Perl 6 and what it looks like when it runs.
**Project Details**
1) The object is for the PL/Perl6 test suite to pass on the latest stable release of Rakudo, or whatever Rakudo devs would prefer that works with PL/Perl6.
Currently, one of the PL/Perl6 tests causes PostgreSQL to coredump due to a change in Parrot or Rakudo or both. This needs to be debugged and either something in PL/Perl6 , Parrot or Rakudo needs to be fixed.
This will involve working with Rakudo core devs as well as Parrot core devs to resolve this situation. More tests will also be added in the process to insure against future breakage.
2) This is actually the most work and most important part of this grant. Without a continuous integration system, something like PL/Perl6 can't stay alive. Since PL/Perl6 ties together multiple pieces of *very* different software (PostgreSQL, Parrot, Rakudo) it needs to continually be tested on new versions of these dependencies to make sure it continues to work properly. Also, this continual testing leaves the data behind that people can look at to fix and maintain PL/Perl6. Without that data, maintenance isn't really possible.
The data in this CI system will provide the data that Parrot, Rakudo and PL/Perl6 developers need to fix bugs and make improvements with the knowledge that things are not being broken.
The needs of the CI system are:
On each new commit to Rakudo nom branch (the development tip), test against:
Postgres 9.1, PL/Perl6 master, g++ compiler
Postgres 8.4, PL/Perl6 master, gcc compiler
On each commit to PL/Perl6 on any branch:
Postgres 9.1, Rakudo nom, g++ compiler
Postgres 8.4, Rakudo nom, gcc compiler
This is just the beginning, but with this, we will have a wealth of data about how versions of Rakudo and PL/Perl6 behave together, across the two major versions of Postgres that PL/Perl6 supports. Also, having data for both a C and a C++ compiler is valuable because many bugs only become apparent in one or the other.
3+4) The user-friendly developer + compile/install documentation will be a Git repo that is also a website, hosted for free on Github. This will also make it easy to get contributors for the website as Github pull requests. This could be hosted in the perl6 Github organization, in which case all current committers to perl6 would have the ability to update the website.
5) Screencasts have become a very important tool for gaining new users and providing entertaining introductions to new subjects. To my knowledge, Perl 6 has never had a screencast, and if it has, it surely can benefit from more. The estimated time of the screencast is 2-5 minutes. It will go through the process of writing a stored procedure in Perl 6, running it, modifying the stored procedure and running it again to see the modified output. If time allows, other examples will be shown and part of the compile/install process will be shown.
**Project Schedule**
I expect about 4 months, working 1/4 time (10 hours per week).
**When can you begin work?**
January.
**Report Schedule**
I plan to provide at least monthly blog posts with updates.
Note that payments on the grant will be contingent on the successful acceptance of milestones and the submission of reports.
**Public Repository**
https://github.com/leto/plparrot
Releases are also hosted on PGXN, which is the Postgres equivalent of CPAN:
http://pgxn.org/dist/plparrot/html/plperl6.html
Where will the change logs be held?
https://github.com/leto/plparrot
**Grant Deliverables ownership/copyright and License Information**
Artistic 2.0
**Who and/or which organization will have ownership (copyright) of the grant deliverables?**
Jonathan "Duke" Leto and other contributors
**Bio**
I have hacked on Parrot Virtual Machine and Perl 6 ever since I attended a Parrot Hackathon in 2008. I started off hacking on the Perl 6 test suite, after getting a commit bit from Larry at the hackathon due to my interest in fixing some bugs and adding tests relating to math functions and complex numbers. I have made contributions to Rakudo Perl 6, the Perl 6 test suite, written Perl 6 modules and I am still active as a Parrot Core Developer, Board Member of Parrot Foundation and Community Manager of Parrot.
As for other credentials, I have a Masters in Mathematics from University of Central Florida and have published various papers on differential equations, as well as being a coauthor of the Google Summer of Code Mentor Manual.
**Amount Requested**
$5000 USD
]]>
Comments (1)
I like the idea and the structure of the grant, but there's one point that I think needs to be addressed.
The nom branch -- which is now the main development branch of Rakudo -- fundamentally changed the way that Rakudo stores objects, and how it interacts with Parrot-level objects.
It now "boxes" the parrot primitive types (int/float/string) into Perl 6 objects, rather than inheriting from them. It can also handle these primitive types directly, but that needs special care on the callee side. Other, non-primitive Parrot objects cannot simply be passed to Perl 6 space, they need to be wrapped into "6model" objects (that's what Rakudo now uses for storing objects).
As a consequence I'm unsure how much work Deliverable 1 is. The proposal text sounds like it is a bug fix plus some updates, but it might be much more, and might even require a redesign of PL/Perl6. Or maybe not -- my knowledge of PL/Perl6 is insufficient to judge that.
That's why I'd ask Jonathan Leto to make a rough sketch of how PL/Perl6 and 6model fit together, and how much work that change will evolve.
