This is one advantage of the use of text files in KiCad rather than binary data
1 Like
Well, I wouldn’t exactly call it “reasonable” to have it outright rejected if I can’t do it myself. I mean, the developers are working together to complete a common plan, not just their own individual ideas. I’m really just making a suggestion for what I think would be a better plan for development.
But are you suggesting that my proposal would be allowed if the contributors really are willing to work on it?
I am not good at C++, so I have to try really hard to do anything, and I still can’t achieve much. Maybe I’ll experiment with attempting to port the reverse rotate feature to 4.0 just to see what I can do. But I will not be able to do anything else as it would take me too long just to do that, and I don’t have the experience to solve any bugs that may come up.
Now this has got to be the most unrealistic suggestion posted on this thread. I have compiled the unstable versions, but as I said, they are just too unstable. I don’t think even the bleeding-edge type people could use it. It just crashes way too much to be of any use. Sometimes, it’s hard to even check out the new features before it goes down.
You’re comment isn’t telling me anything I don’t know. It’s nonconstructive, and it ignores the points that I’m making.
Then help to improve the documentation, fix symbol and footprint libraries, confirm bug reports etc. All are valuable and let the coders get on with coding
1 Like
Sorry mate, I haven’t had the need to compile anything yet. I’m under Windows.
And I use nightlies 100%, I just don’t change the horse while I’m in the middle of a project, nor is my work so important that I would need to rely on stable releases.
Currently I’m running nightlies from around February which are - knocks on wood - stable and work. No crashes or anything.
But I wouldn’t do the same with the nightlies from March really 
Obviously your mileage varies from mine, thus the difference in opinion.
So if you can, check out the source from start of February, compile it and see how you go.
Still non-constructive?
3 Likes
I don’t know, I am not the project leader nor a mind reader! I am quite certain from my background in software that a minor release now would simply delay the future 5.0 release.
The purpose of nightlies is to find crashes and bugs, not to provide an advanced release for users. There seems to be a widespread misunderstanding about that. If the nightly crashes then it is meeting its intended purpose : finding bugs before the software is released. If you really want to help, investigate the crash and post a bug report.
To be fair, you are ignoring most of the important points that others are making, particularly debugging is difficult and KiCad has very limited resources.
As I also mentioned, you would need to present a persuasive case to Wayne, I don’t think he even visits here, so I regard this whole thread as non-constructive. Sorry!
3 Likes
yeah i would say right now that a 4.1 definately not coming from master any time soon, as some nightly boards are not compatible with version 4.0 of kicad, and you don’t break that stuff inside a minor release. There have also been some rather large deep changes that would make backporting features a pain in the ass and just not worth it
3 Likes
My work situation is similar: neither so complex, nor critical, that a proven fail-safe tool is demanded. I’m running a nightly version from around 7 - 10 February, on three different Win7/64 machines. It seems to crash once or twice a week, but I can’t put my finger on the exact circumstances. The misbehavior may be confined to one machine in particular, so it may not be KiCAD’s fault. All I lose in the crash is the work in the buffer since the last SAVE.
For a while I was grabbing nightly builds that were a day or two old. (If no complaints appeared here on the Forum in the first day or so after release, I judged the nightly build safe enough to try.) Around mid-February the nightly builds started getting buggy and unstable, and one of them corrupted some library files. That’s when I reverted back to the version I’m running now.
About a year ago I downloaded sources and compiled under Windows just to see how the process worked. It was a couple orders of magnitude more complicated, and more aggravating, than simply grabbing a pre-built installation package from the nightly downloads at http://downloads.kicad.org/windows/nightly/ .
Dale
2 Likes
Yeah, but we haven’t really had a policy that covered minor releases since we…haven’t done any. I would be very strongly against breaking any compatibility in any way in a minor release.
and just because there is a policy in place to not be forward compatible i would be interested to see an example of software that broke stuff in a minor (not including boost which is well… boost)
The rounded rectangular pad feature was introduced into Nigthlies ages ago.
These are actually preferred in the coming IPC-7351 version C
is there even an eta on the release of that yet?
Sure - later this year, probably. 
According to Stambaugh’s FOSDEM talk the plan was stable release sometime in summer. Which sounds overly optimistic at this point.
The estimates are always overly optimistic. I’ve been saying Q4 2017 for a while, which seems more realistic to me.
We don’t have that many things to go on the 5.0 roadmap, I doubt it’ll be pushed back much further than that.
I did say late last year, that a 4.1 would be a good idea, but was shot down as 5.0 was impending.
This was before the heavy lifting in the code base started, so forking would have been much simpler then than now.
Two years is a long time between releases.
Meanwhile todays Ubuntu Nightly is 100% non functional for me as Cvpcb crashes reliably, yes I have registered the bug 
1 Like
You got unlucky, as the nightly was built from the commit just before this was fixed. Try the next one
What do you think would be the “right time” between releases?
The last release process took 9 months (from code freeze to stable 4.0.0 rc2 release date).
The right time is when there is enough feature change for most users to notice.
If stabilisation takes 9 months again, we are looking at late 2018, by which time another Ubuntu LTS will be around with newer Boost and wxWidgets to make life more complicated
1 Like
Personally, I doubt stabilization will take nine months again. The 4.0 release had a huge amount of new code that kept having new bugs found and delaying the release…5.0 does not have nearly as much.
Then again, the 4.0 release kept surprising me with how far it could be delayed, soooo
1 Like
Weird thing is, in all the projects I’ve known, regardless of the estimated timescale there has never been one that released earlier than expected. Usually what happens is people think “oh, we are ahead of plan, let’s add some more features”.
Anyway, I was expecting Wayne to announce a design freeze already, any idea when that is likely to happen?
1 Like