Forum Archive

I really just want OpenCV and Scipy.

Scipy support makes sense and doesn't seem like it would be much harder to incorporate than numpy was. Scipy would also mean pythonista had all the necessary dependencies for scikit-learn, and we could have machine learning in pythonista.

OpenCV would make Pythonista SO much more powerful for building apps. Document scanners, augmented reality, and video processing would all become possible with OpenCV. I'm not sure if it's possible to download OpenCV and use it with objc_util, but I'm still scared of objc_util as of right now.

@Cethric is doing great work implementing OpenGL, but I'm not even sure objc_util will make it past apple's review guidelines.

I think this brings up the topic of dynamic libraries and linking again. Porting these two libraries or potentially solving the issue of supporting both Python 2 and 3 could be made much more "doable" by packing the functionality into a dynamically loadable framework. It seems to be possible now as part of the whole App Extension capability but it is really hard to tell. I spent a bumch of time on stackoverflow reading about the topic and can't find a single example of a project yet that takes it all the way through to publishing an App out to the store. I just find lots of people having all kinds of build issues and then claiming that it worked finally with a list of caveats (like only on IOS8.x, etc.). Here is one sample thread:

http://stackoverflow.com/questions/25909870/xcode-6-and-embedded-frameworks-only-supported-in-ios8?lq=1

Yes - I know this is talking about "embedded frameworks", but this appears to be a form of dynamic framework. It just adds to the confusion for me.

That's way past what I understand, but interesting nonetheless.

I wanted to register my concern about the super long beta cycle for Pythonista v1.6.

In these days of continuous delivery it worries me that there have not been more upgrades that were smaller in size but more frequent. Core Bluetooth could have been an upgrade. Ctypes could have been an upgrade. ObjC_util could have been an upgrade. Pillow could have been an upgrade. IOS9 (appex, etc.) compatibility could have been an upgrade. Today all of these capabilities are only available to those who are enrolled in the Beta program. More frequent upgrades would mean that we would be at Pythonista v2.5 instead of v1.5 but a bunch of users would have access to these incredible capabilities. I would like to vote for more releases with less functionality per release as opposed to the long beta cycles with tons of features that we currently have.

Agree. I think 1.6 could've been a UI upgrade, and other releases could've focused on adding appex, adding more modules, etc. all in separate releases. I'd be interested @omz in hearing the rationale behind the one big release? Is it just that you start working on new features before fixing bugs so it's never really stable enough for a release?

I'd be all about smaller realeses more often. I've been waiting for a long time for some of the features I've seen talked about on the forum. I've put myself on the list for the beta multiple times, but to no avail.

I think he only adds people to the beta when he releases a new version of the beta. He hadn't done so in a while.

On the topic of releasing features in smaller updates - I think that part of omz's reason for keeping features in the beta for a while is that they can still be removed safely if necessary, without upsetting people who already used them. (So far this has happened for Apple Watch and iCloud Drive support, both were in the beta for a short while and were removed again after a few builds.) Of course it's never nice when a feature is removed, but users of the release version will have different expectations than beta users. ("I paid $6 for this, and my favorite feature gets removed? Unacceptable, I want my money back!!!!")

Yes, but this isn't entirely unavoidable. Open in functionality was removed in 1.5, features will invitably have to be removed outside of the beta occasionally.

Seems like submitting smaller reviews would be a faster way to find out what makes it past review. Then again, if a feature that Apple might not like got buried among a huge amount of other changes... I guess I can see an argument for larger releases. ;-)

Sk module was also put in a beta but removed.

Yeah. Never got to play with that one.

@ccc The sk module has been added back in. This may have changed, but I looked in the Standard Library/site-packages/sk.py and omz's NotImplementedErro() was removed, and I was able to play the word game. Some of the functionality has been removed, so the platformer game won't work, but shaders do.

Cool. Looks as if documentation and the built-in editor are still gone.