On 07/12/13 00:56, Benjamin K. Stuhl wrote:
Why not keep the document tree format the same python text that it is
now and just store it as a string in the HDF5 file? The only objects in
the save file that are large enough to benefit from HDF5 storage are the
datasets themselves, and adding a new "create Veusz dataset from a
dataset in this HDF5 file" command is easy enough. So the hypothetical
.v5z binary format might be an HDF5 structure with
/veusz -- the .vsz stored as a string
/datasets/* -- actual storage for the datasets
Ok, I thought it would have to be stored as an attribute, which is
limited to 64K, but it looks like a variable-length string dataset would
be ok.
I quite liked the idea of a JSON/XML/YAML saved file option for veusz,
as this would be easy to manipulate outside the program, but I can think
about this later.
Jeremy
_______________________________________________
Veusz-discuss mailing list
Veusz-discuss@gna.org
https://mail.gna.org/listinfo/veusz-discuss
