Custom I/O?

[TJA]

I've just started to integrate Alembic into my program, and so far, everything works fine and is easy to use. However, it seems that the archives always require a filename, which means that I need to touch the disk every time I want to read/write Alembic. I have a few use cases where I pack multiple abc files into a single one, and have them in-memory only, so I would like to load directly from a buffer or alternatively store directly to a buffer. Is it possible/planned to provide custom istream/ostream class (or another stream abstraction -- as long as it's simple) into the archives which are then forwarded through HDF5? From a quick glance at the HDF5 documentation, it seems that they support custom I/O routines through the "Virtual File Layer". Or is there some other workaround to avoid touching the disk when reading/writing alembic data?

Cheers,

  --t

[Lucas Miller]

We are considering it, but it isn't a task that is formally on our list yet.

If you happen to be on a Unix-like system, you may have access to /dev/shm which would allow you to read and write  to memory.

Lucas

--
You received this message because you are subscribed to the Google
Groups "alembic-discussion" group.
To post to this group, send email to alembic-d...@googlegroups.com
To unsubscribe from this group, send email to
alembic-discuss...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/alembic-discussion?hl=en
 
For RSS or Atom feeds related to Alembic, see:
 
http://groups.google.com/group/alembic-dev/feeds
 
http://groups.google.com/group/alembic-discussion/feeds

[Timothy J. Arrington]

Allright, good to know that its being considered. At least for my
use-cases, this would be a great advantage (also makes unit-testing
painless!)

Unfortunately, I have to support both Unix and Windows, and that's
part of the reason to get my custom memory streams into Abc.

Cheers,
-t

2012/5/25 Lucas Miller <miller...@gmail.com>: