>  >    Patrick, you may want to start working against CVS HEAD, as it 
>  > will make everybody's life a little easier when it comes down to merging
>  > patches/changes.  Thanks.
>  > 
> 
> You can now get patches against CVS HEAD:
> http://www.chbcp.net/~pat/geda/libgeda.action.diff.gz
> http://www.chbcp.net/~pat/geda/gschem.action.diff.gz
> 
> Regards,

I've just taken a look though those diffs.

Their scope is far wider than ever implied by the topic at hand, and
unfortunately impacts a great deal of the noscreen changes I've been
working on. As far as I can tell (without spending a day or so looking),
it would be incompatible with most of the tried and tested work on
glist-dev.

The intent of the patches look good, refactoring stuff which should be
in libgeda out of gschem, re-writing functions to tidy them up, however
many of these functions are key to the noscreen changes. Where they
needed re-writing, or grouping together, (as part of the noscreen
changes), I did that.

I have also (separatly in git) made use of cached world bounds (instead
of screen bounds), to speed up hit detection, moved bounds recalculation
of bounds out of the ..._get_bounds() functions, and into
the ..._recalc() functions.

On 23rd Nov, I emailed the list to say I was looking into noscreen
changes. I appreciate Patrick, your changes will have taken some time,
and may even predate me starting the noscreen experiment, but it ought
to have been obvious that we were both working in the same area from my
emails. There have been emails discussing branching for some development
work in this area, (and I've been waiting for glist-dev to merge for
that).

If you'd mentioned what your work covered, I would not have put so much
time and effort into my own. (Instead waiting to start from a cleaner
code-base). I don't see any email indicating your work until 6th Dec.

As far as I know, Carlos would also be unaware that you were working on
alternative code to that he has devoted his time to producing (in open
view of the other developers).

GAF and PCB rely on people devoting their free time to coding. I enjoy
doing so, however I don't enjoy having to needlessly duplicate effort
for the want of communication. If I was being _paid_ to code on gschem /
PCB, I'd still be irritated. It is problems like this which can put off
developers, which can't be good for the project.

It has been my intent with the noscreen changes, that the code-base
should remain as close to its current architecture as possible, such
that testing of rendering speed changes could be made sensibly.

As Carlos's glist-dev work has proven to be stable, I think it ought to
be merged to HEAD, and future development work based on that. Any future
modifications can just as easily be based upon that codebase.

The noscreen work / experiment I've done, I think has some useful code.
Not all of it is purely noscreen related, however it the later (less
noscreen related changes) are all separate, and very small commits, and
are made possible / sensible / necessary by the noscreen changes.

Collectively, they simplify the codebase significantly. It isn't
complete yet, but I think represents a cleaner platform to start future
refactoring work.

I'm busy with work this week, so can't rush to re-base my trees
incorporating the latest merge and changes to HEAD, nor even begin to
contemplate integrating Patricks changes.

As I have always intended to do with my noscreen work, is it possible
Patrick, to produce smaller patches addressing individual issues? Can
you fix the moving / copying without touching so much other code?

I'm looking forward to having time to setup gitweb, so people can see
the small commits addressing different issues I've spoken about.

Regards,

Peter Clifton.




_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to