[L2Ork-dev] cyclone merge resolution (Re: updating cyclone)

Alexandre Torres Porres porres at gmail.com
Wed Jun 14 19:48:25 UTC 2017


Hi, thanks for doing this, I appreciate it.

For those not aware, this has been a long discussion, with many turns and
lots of new information to me. Let me summarize.

When trying to merge cyclone, the libloader failed, cause we had both a
path and a library with the same name to load. As I see it, the options
here are:
1- fix the loader
2- get rid of the library for purr data (we can have a workaround with
hexloader, and this would also mean a separate branch)
3- make cyclone into a single library named cyclone
4- rename the library

But why do we have now both a path and a library? The library is for
loading operators with non-alphanumeric characters, an issue that is only
solvable in vanilla with a library, or a third party external dependency
(hexloader). The reason we have both it and a set of single binaries is
that we inherited the single binaries from the last maintenance phase.

If I'm not mistaken, cyclone used to come as a single library binary, but
got chopped off into separate binaries in Pd Extended because of some
Pd-Extended issue I don't know about.

So, on to my opinion on the solutions. I've listed them in order of
personal preference, kinda...

Fixing the loader is option number 1, it would introduce the least change,
and would bring benefits of improving the loader system.

A second option would be to create a branch in Purr Data that just loads
files with hexloader (like it is doing with zexy). This is actually
something jonathan proposed first. This forces the creation of a separate
branch for the Purr Data help files and all, but I don't think it is that
much of a hassle and I could work on that, no problem.

The most complicated issue, on my opinion, is thinking about what to do
with the "operators library". This just opens a can of worms in my head...
and I'm sorry if it is hard for me to take a decision about it, I'm just
not sure which is the best one.

One issue is that I already forced a name change. In cyclone 0.2 it was
called 'nettles', I hated that name to the guts (I never knew what it
meant, it was never part of cyclone before, I felt it was hard to relate
that name to cyclone), so I thought it was best to revert back to the
original "cyclone" library (which used to call these objects).

I actually also have an issue on the need of these two part deal, loading
cyclone both as a set of separate binaries, where you need to include
cyclone's path, but also a library only for a subset of objects. That was
never a good design decision. It would be preferable if it'd be just one
thing, either all separate binaries, or just a single binary pack.

Again, vanilla can't load non-alphanumeric, so that's out. I'm dreaming we
could try and make this into vanilla, but I should get real...

and well, I never discarded the idea of making cyclone into a single binary
pack. That would make it simpler, cleaner, and would also fix the loading
issue in Purr Data. So it feels like a good option.

Renaming the library yet once again is something that would be a last
resource kind of thing for me. In an outburst of impulse and anxiety, I did
propose and changed the library name to 'cyclops' (standing for 'cyclone
operators'), but jonathan felt this could potentially cause name clashes
with user defined abstractions and proposes "cyclone_ops" - one can also
consider going back to 'nettles' (yikes). One way or another, again,
touching the cyclone name is the most delicate issue to me, and it's what
gives me more headaches.

So, yeah, I'd love to hear more opinions, please Derek and Matt, share your
cents, and anybody else on this list. I'll keep thinking about this in the
meantime.

cheers

2017-06-14 15:56 GMT-03:00 Jonathan Wilkes <jancsika at yahoo.com>:

> Actually, let me clarify to avoid a digression--
>
> This vote isn't about whether pd-cyclone should accept the pull request.
>
> This vote is about whether Purr Data should merge pd-cyclone with the
> "cyclops" binary name, or
> the one from the revision with the "cyclone_ops" name.
>
> Thanks,
> Jonathan
>
>
> ------------------------------
> *From:* Jonathan Wilkes <jancsika at yahoo.com>
> *To:* Alexandre Torres Porres <porres at gmail.com>
> *Cc:* Derek Kwan <derek.x.kwan at gmail.com>; "l2ork-dev at disis.music.vt.edu"
> <l2ork-dev at disis.music.vt.edu>; "brbrofsvl at gmail.com" <brbrofsvl at gmail.com
> >
> *Sent:* Wednesday, June 14, 2017 2:51 PM
> *Subject:* cyclone merge resolution (Re: [L2Ork-dev] updating cyclone)
>
> Hi everyone,
> I've got a pull request open on pd-cyclone to change the name of the
> binary for loading the ops.
>
> You can read about it here:
> https://github.com/porres/pd-cyclone/pulls
>
> I want to know if it should be accepted before I merge pd-cyclone into
> Purr Data and make the next release.
>
> Albert, Ico, Alexandre, Matt, Derek? How do you vote?
>
> I abstain.
>
> -Jonathan
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://disis.music.vt.edu/pipermail/l2ork-dev/attachments/20170614/3bde0d79/attachment-0001.html>


More information about the L2Ork-dev mailing list