Jonathan Steffan wrote:
Karsten Wade wrote:
> ----- "Jonathan Steffan" <jon(a)fedoraunity.org> wrote:
>> It's still in the works. I've had some setbacks moving from Plone
>> 2.5.3
>> to Plone 3 (as a Plone developer). I'll try my best to put together
>> something for beta testing but I'd really like to look at some other
>> unrelated changes Ive been playing with (mainly getting rid of the
>> Makefiles). I'll try to put together an email with a score of
>> suggestions to test the waters for where we could go with this. Right
>> now the buildd (the daemon that interacts between plone and cvs) uses
>> the existing Makefiles but requires them to be error free and also
>> requires the innermost Makefile to be valid for building the
>> document.
>> It's very fragile to say the least. I'm working out a pure python
>> based
>> build system that will replace the Makefiles with simple/nice config
>> files for each module, among other things. I'll try to send this email
>> soon.
> Recommend that you send that email *before* you do any coding. I'm not sure what
is fragile to Plone about the Makefiles,
> but it sounds like you are suggesting to duplicate their actions
entirely within Plone.
Actually, the changes I would like to make will not have anything to do
with plone. Plone will just be able to trigger actions, in the same way
a human would.
Thereby making a duplicate, parallel system to understand and maintain.
Well, this is already the case. There is static logic to find the
innermost Makefile for the buildd (read: plone action) to actually work
with
Hmm... thunderbird--
This already being the case:
Changes to the way the Makefiles work would require changes in the logic
of the buildd, albeit only a few functions.
> Unless you plan on personally porting changes from the Makefile into this new system
$FOREVER ... it doesn't seem like a plan that can scale.
No, I'd like to see a pure python based solution replace the Makefiles.
We have more in house python skills then any other; IMHO.
> We already are resource challenged taking care of just one toolchain.
Yes. I know. My thoughts on getting rid of the Makefiles has been from
observing comments about how much of a PITA they are to work with.
> If the Makefiles in CVS have errors or are invalid, that is a larger problem than
just Plone.
> Wouldn't the better solution be to fix the central Makefiles?
I was proposing replacing the Makefiles.
Jonathan Steffan
daMaestro