On Sat, Jan 26, 2013 at 7:07 PM, Peter Stuge <[email protected]> wrote:
> But openocd.git is not the place for proof of concept code,
> it only includes finished work - so that it can be released
> at any time.

Not sure if there is any "finished work" anywhere, not even in the
openocd.git. My code is not a dirty hack, its clean working good
starting point. There must be some starting point, and the progress,
then some goals, then further developments, etc..

I can see some contradictions in your argument - you want the code to
be finished, you want small changes, you want people to test it, you
don't want to integrate it, etc. I don't know where this will lead...

I like that you are not accepting all random patches and want to keep
the organization clean, however I thouhgt GIT is for the development,
that is expanding software step by step, not just keeping the release
code.

What is the other way to add such big changes to the project when the
change is ready?

Maybe, if you don't trust my changes (which is reasonable assumption),
two of you as the most advanced developers should simply sit on the
changes, have some time (one month lets say) to analyze it line by
line, make a problems list, make a todo list. And then we can discuss
over the code. I would prefer to talk over the code rather on the
ideas on what thinkgs should be because most ideas are simple in the
head and hard to code being restricted with existing design...


>> It would be nice to get some support at this point.
>
> Yes, I think that's very reasonable, and well deserved. But as you
> know, many if not all OpenOCD contributors are busy, which makes it
> very difficult to process a huge set of commits - as opposed to a
> small set.

This is one functional change - no matter is there is more smaller
patches that change file by file (as you like?), or less patches that
group the changes (as I like) - the result will be exactly the same.
This is why I suggested to concentrate on the core development rather
than patch editing. The change is big, there must be more than one
line of the code. Please note that I only changed one thing in
existing structure of OpenOCD and that thing was experimental anyway.


>> If anyone thinks that some things can be done better, feel free to
>> code and show the ideas working.
>
> On one hand you're perfectly right, on the other hand it is often
> significantly more efficient to discuss and think about ideas than
> to implement and test, with the minor catch that discussion requires
> working communication.

This is why I am communicating my will to change internals and the way
I like it to be done (separate thread), there is a time for discussion
and argument between different point of view. The goal is to create a
better tool.

I prefer to talk over the code, because anyone can write anything on
the list, this takes time to be written and then read, while this time
also can be used to create code that will prove the idea, otherwise
its just trolling. If the code cannot be created easily then the idea
maybe needs to be reconsidered. I spent lots of time to invent and
implement ideas that will work and does not change existing OpenOCD
structure - If things were solved that or another way, there is a
reason behind that - the most probable reason is non-invasive
approach, backward-compatibility and elegant solution. If things were
solved that or another way, that's probably the other way did not
bring the final solution any better. Writing the code yourself will
show that and I want you to understand that :-)

I guess you want to make too much changes at once. Lets divide the
work into steps and do it step by step. We cannot do all of the stuff
at once in the Gerrit. I really think Gerrit should be just a filter
for patches and some buffer so MINOR changes can be introduced (i.e.
typos fixes, etc). Gerrit is not a GIT, its not branch, but you
already know how I see this - you shifted GIT/MASTER/HEAD function
info Gerrit which I don't really like. If the patches are not as
supposed to be, they should be simply rejected and sent again in
another form. I see Gerrit as patch cleaner and review tool, not the
development branch of the project. Here is our disagreement.

Btw. what are the exact problems with my patches/commits, except
commits needs to be reordered (which makes no difference to me)? I did
not get any precise list of remarks to the code, and the ones showed
up on the list are totally out of the scope of my changes. This is why
I am just waiting to see what was wrong and what was the solution ;-)

Best regards :-)
Tomek

-- 
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info

------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnnow-d2d
_______________________________________________
OpenOCD-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openocd-devel

Reply via email to