Tragic Lantern for 6D

Started by 1%, December 24, 2012, 07:07:02 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

sam.anstis

I've had a few crashes recently with the magiclantern-Tragic.2014Jan18.6D113.zip build. Whenever I load ML, it says the camera did not shut down correctly, skipping module loading.

Here's a few crash logs.

https://www.mediafire.com/?gsdipqqecsdeqhl
https://www.mediafire.com/?pkx3hexqj96d7rx
https://www.mediafire.com/?gb0tdcgv1fetxe6
https://www.mediafire.com/?db6qqsb6npmlhz3

1%

Memory

at ./Memory/Memory.c:568, task RscMgr

try loading less modules.

sam.anstis

Quote from: 1% on February 09, 2014, 05:10:52 AM
Memory

at ./Memory/Memory.c:568, task RscMgr

try loading less modules.

Problem solved, thanks

D6-2035

Sorry if this is a stupid question.

magiclantern-Tragic.2014Feb08.6D113.zip?

What has changed to the magiclantern-Tragic.2014Jan18.6D113.zip version?

1%

a whole ton of commits. read through the repo.

Freemountain

Reason for white screen when fpsoverride used? No, not the too long exposure;)

D6-2035

Quote from: 1% on February 09, 2014, 07:26:07 PM
a whole ton of commits. read through the repo.

OK, will do, thank you.

RTLdan

Hi all,
Just a little confused over the whole TL vs the new 6D ML builds.
1% has done an awesome job for the 6D and I've been really getting so much use out
of TL while ML's version has been "unmaintained". Everything has been incredibly stable and very up to date with new features. I guess what I'm wondering is, will I be losing some of the features I've come to expect from TL if I begin using builds from ML? I want to respect the "unity" that ML is asking for by using those builds, but on the other hand if no one is really working on the ML 6D build, it's asking a lot to sacrifice all the work, testing, and features for my 6D that 1% has already provided.

Really just not sure what the story is or what I'm supposed to do?
Thanks to all the developers for making any of this even possible!
-RTLdan

1%

I think the whole deal will be to take what is really stable and port it back to main so code there is actually up to date not 4 months behind. Plus I get to see if bugs are related to me or not and everything remains coherent and together. Plus get some code review from better programmers and things end up being more efficiently done.

a1ex

Well, not exactly. The community split caused by Tragic Lantern only serves to cause confusion, and since it diverged a lot from the main builds (and keeps diverging), merging back is getting more an more difficult (almost easier to code these things from scratch).

For me, the lack of a positive answer from you, 1%, for more than one month in our private discussions, more than 4 months in this public discussion and more than one year since Tragic Lantern exists counts as refusing to collaborate (and just talking is not enough; I've expected you to take some action). I'd prefer to avoid making use of my admin superpowers to prevent further damage to the ML community, but it looks I have no other choice.

Quotewill I be losing some of the features I've come to expect from TL if I begin using builds from ML?
Since we have done a blind backporting (without a 6D in our hands), you are the one who should answer this question and tell us what else to backport: http://www.magiclantern.fm/forum/index.php?topic=3904.msg101892#msg101892

You should be aware that 90% of the features from Tragic Lantern come actually from Magic Lantern (some features need fine-tuning for each camera). There are some things implemented in unclean ways (mostly bitrate control), but these are going to be cleaned up at some point and merged back once we have a solid port with the basics working really well.

QuoteEverything has been incredibly stable and very up to date with new features.
From my code review attempts while trying to backport the changes, and from the feedback from power users (e.g. Marsu42), I wouldn't call it incredibly stable. I've tried to do my best in disabling the unstable bits (especially the code that - instead of fixing the root of the problems - simply disabled the error message in either Canon or ML code), but I have no 6D to test the changes on (so your detailed feedback is essential).

http://www.magiclantern.fm/forum/index.php?topic=9510.msg78084#msg78084

http://www.magiclantern.fm/forum/index.php?topic=7503.msg83832#msg83832

http://www.magiclantern.fm/forum/index.php?topic=8900

http://www.magiclantern.fm/forum/index.php?topic=9508.msg101507#msg101507

More details coming soon.

Marsu42

Quote from: 1% on February 16, 2014, 08:02:26 AM
I think the whole deal will be to take what is really stable

If it would be you managing to submit dedicated backport pull requests for code review & integration, great, and you indicated such before... but if it's "I'm developing my own fork as it's most convenient, but feel free to dissect it and copy/paste to ML" this won't work, and you know it.

It's perfectly legit to do it your own way of course, and I'm absolutely grateful for all the 6d work you've done as I'm using it right now - but as time passes & the more the TL code base differs, the more pressing is the need for you to decide for one approach or another.

The only way I can currently see to save this is for you to start submitting small pull requests once you update new ML features to the 7d/600d/6d, like stub addresses or such - so you'd help "maintaining" these for ML and don't just keep them in your fork. If that's covered, surely a schedule could be worked out what features need to reach what stability for a backport pull requests, of course it's fine to work on unstable features in an experimental fork for some time.

However, unless you indicate with action you're going to be a part of the *ML* community, as a long-time ML user I can only side with alex and say the split in the community merits stopping users being confused, i.e. separate ML & TL entirely.

I know maintaining an own fork is much more fun, submitting pull requests is tedious and having them torn apart by smart-a** comments by the big devs can be very annoying as you're not in control of what gets merged. But I do hope that if you should decide to take up the offer to make a schedule for TL re-integration, I can only hope the ML devs will acknowledge that this might not be easy for you and will strive to let you put your ideas into ML - all factual disputes I've ever read about can be mediated or solved quickly.

RTLdan

Hi guys,
Sorry for my earlier post... I didn't mean to stir anything up, just really wanted to know what the difference was and what was recommended to do for 6D owners.
I should have clarified too that when I said "incredibly stable", I meant that the features I use Magic Lantern for (mostly the meters and intervalometer so far) it has been reliable. No crashes to report. I understand there may be instability in some of the newer features that I have not been using much.

Anyhow, I really do appreciate all the work that every one of you developers have done for the community, and while I'm not a coder, I'll download the ML 6D nightly and test it as best I can to help with the development. As an outside voice though, I just want to say that I'm glad there is someone like 1% with a 6D AND coding skills to keep this camera in the mix. It's been a blessing to me to have access to these amazing features.

Thanks again everyone!
I'll watch these threads closely.
-RTLDan


sparedog

So I have been using ettr for quite while now and usually it works really well. However, I must have done something as it is not behaving itself.
When I try to ettr in manual mode, ML only tries to change the iso settings, and nothing else. Before.
, it seemed to change the shutter settings first and if it couldnt get within range, it would then change iso. now it just changes iso, no matter what I do or what the camera is set to exposure wise.
I have even tried to test it by manually ettr ing and then auto ettring, only to watch the camera ramp up the iso until the max it is allowed to do and coming up with an error.
Has anyone had this experience and know the fix?

1%

Are you sure auto expo isn't on too?

bjacklee

my 6d stops responding everytime I try to preview the raw image and try to adjust the exposure by holding set btn while scrolling the wheels. anyone has thesame issue?

1%

Works fine as of todays compile... might be issues from conflicting shortcuts in canon or more left over stuff from "new lv buffer detection"

hagar-dunor

Hi all.

I'm a new 6D user and tempted to install TL. First of all, kudos to 1% and others working on the 6D fork, probably noone would deny the great work done here.

But I haven't installed TL for now, and hesitating to do so mainly because of the unclear future of TL. To me, a community supported version merged to ML is far more important than having bleeding edge features. 1% you're very reactive to fix bugs or integrate new features, why not use this workforce to freeze TL for a while and merge it ? For what it's worth, I would be ready to donate if it can make this merge happen. On the other hand, I won't donate for TL if it continues to diverge, one could wonder what's the TL goal actually.

Cheers

Marsu42

Quote from: bjacklee on February 28, 2014, 09:58:00 PM
my 6d stops responding everytime I try to preview the raw image and try to adjust the exposure by holding set btn while scrolling the wheels. anyone has thesame issue?

Same over here, problem disappeared after switching to ML.

Quote from: hagar-dunor on March 03, 2014, 02:48:04 PM
1% you're very reactive to fix bugs or integrate new features, why not use this workforce to freeze TL for a while and merge it ?

Amen - though looking at the commits, the merge seems to be on a better track now.

1%

Its working here, I haven't been able to reproduce it. Do you have the set  button mapped to something else? I'm merging 7D at the moment + a lot of the stuff in 6D is already merged and the rest is not accepted in main (i.e cache hacks). What are you missing in ML? Have some bugs?

QuoteOn the other hand, I won't donate for TL if it continues to diverge, one could wonder what's the TL goal actually.

I don't take money donations. ML takes BTC donations. No workforce, its just me and 5 bodies. My goal is a fully featured camera that I can use, its never been much more than that.


bjacklee

Quote from: Marsu42 on March 03, 2014, 03:01:51 PM
Same over here, problem disappeared after switching to ML.

This issue is now fixed on the latest build. :)

@1%, Whats the status of the Digic 5 issue with low fps video? it is still very buggy everytime I use this feature in 6d. but with my 60d, everything works fine.

sparedog

Weirdly, I just did a professional shoot and not one single problem with ettr, it worked every time, and I don't think that I've changed any settings. I will write everything down if it starts working strangely again.

sparedog

Quote from: 1% on March 03, 2014, 05:31:55 PM

I don't take money donations. ML takes BTC donations. No workforce, its just me and 5 bodies. My goal is a fully featured camera that I can use, its never been much more than that.

I, for one, think you are doing an amazing job on this 1%, and dont really see why you should change what you're doing.

Marsu42

Btw 1.1.4 is out. Yawn. I'd advise not to upgrade until some people figure out if there are hidden regressions like Canon blocking 3rd party batteries like on newer 5d3 fw.

http://usa.canon.com/cusa/professional/products/professional_cameras/digital_slr_cameras/eos_6d#DriversAndSoftware