gEDA-dev: [RFC] Sort before write

[Peter TB Brett]

On Friday 19 October 2007 04:19:59 Ales Hvezda wrote:
> > Hmm, I'd better hurry up and get the "sort before write" code into
> > libgeda, then. :)
>
> "Sort before write"?  Is this going to change the order of objects as they
> being written to disk?  Are we really sure this is a desirable change?
> Some people might be a little sensitive to such a change.
>

What:

Sort all elements by some function of position before writing to disk.  
libgeda would be able to read unsorted files, and unchanged files would not 
be written out (as you'd expect).  The exact comparison function used isn't 
important, as long as semantically identical schematics have identical 
on-disk representation.

Why:

1.  The most reliable way to detect whether a file has changed is to use a 
checksum.  At the moment, two otherwise identical schematics may have their 
elements in different orders on disk and thus have different checksums.  
Identical gEDA-format files having identical checksums is a requirement for 
getting the new embedding system to correctly detect when symbols are in need 
of updating.

2.  Changes to a particular area of the schematic will appear together with 
the unchanged elements of that area when viewed as a diff.


I really can't see any conceivable reason why there'd be any objection to this 
behaviour change!

                             Peter
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://www.seul.org/pipermail/geda-dev/attachments/20071019/68b9a170/attachment.pgp