That’s probably because many of the changes are in the binary blobs. They won’t show up in version control.
I did a diff (it’s complicated, because java applications are used as well, there is a lot of “other” software in there.) for fp2-sibon-2.0.0-blobs.tgz and fp2-sibon-2.0.1-blobs.tgz. Nothing has changed. I’m too lazy to compare fp2-sibon-2.0.3-blobs.tgz(?), because all I could possibly find out is that now “something has changed”. And I’m a bit tired of playing Sherlock Holmes … it’s not making things better.
I guess all one can do is download the fp2-sibon-2.0.3-blobs.tgz and compile again (There are things getting linked …).
I have another question:
What makes you think only changes starting with “FP2” are relevant?
I don’t have a clue how Fairphone handles commit messages. However I know that a very common model is to start each commit with the bug ID in the bug tracker to mark that bug as resolved. Of course this only is the case if the code is actually submitted as a single commit and written by Fairphone staff.
But there are other scenarios in which stuff gets fixed but the bug ID is not the first element in the commit message. Maybe bug fixes are merged from another branch (then the message likely starts with “Merge”) or commits from elsewhere are cherry-picked (then the message might be totally different).
Long story short: I’d consider all changes as part of the release and not only those which start with “FP2-”.
Edit: For clarification, you’d have to ask Fairphone staff managing the repository and/or writing code for Fairphone 2. They’ll know.
I have no clue, I just looked for the usual suspects (CVE, FPII, FP2) that normally contain “new” stuff.
You are right that all changes are part of a new release. All I want to know is: Is the code that is currently offered in the repo (the stuff you get with repo sync) code from the 1.2.8 release? I don’t think so, but I don’t know for sure, so I ask the community.
I stoppend asking the Fairphone staff, they are too busy. That’s why we have a repo so we don’t need them if they are busy doing their main job (making new builds for all FP users). The code here is just a very cool extra.
$ repo forall -c git log | grep -i 2016 Date: Tue Jan 5 17:14:27 2016 +0800 Date: Mon Jan 18 18:26:40 2016 +0000 Date: Fri Jan 15 16:02:56 2016 +0800 Date: Thu Jan 7 11:20:46 2016 +0800 commit cb137fc11dcf87be7592016eeab44d924d6227b2 Date: Mon Jan 11 16:06:19 2016 +0000 Date: Mon Jan 11 16:15:55 2016 +0000 Date: Wed Jan 6 17:20:05 2016 +0800 Date: Tue Jan 19 09:56:34 2016 +0800 Date: Wed Jan 13 17:48:02 2016 +0800 Date: Tue Jan 12 15:11:43 2016 +0800 Date: Mon Jan 11 14:36:18 2016 +0800 Date: Mon Feb 29 17:30:22 2016 +0100 Date: Tue Jan 5 15:39:26 2016 +0800 Date: Mon Jan 18 14:22:51 2016 +0000 Date: Tue Jan 12 11:30:16 2016 +0800 Date: Wed Jan 13 18:08:51 2016 +0800 Date: Wed Jan 6 18:33:28 2016 +0800 Date: Thu Jan 14 15:16:34 2016 +0800 Date: Mon Jan 11 14:34:25 2016 +0800
Update: TLDR: Is there a way to get the “new” code from the repo that was used in the new build?
I spend time searching around. I think the new branch could be
fp2-sibon-2.0.2. I will sync and report back.
This approach is also flawed because commit times are kept even if a commit is cherry picked from another branch. That means, is someone (e.g. at Google) has committed something last year but the Fairphone team only took over that code this year, it would still show the old date.
I don’t know of a good way either. But the tree view of git is already sorted by time of addition of a commit (to this particular branch). That means you can just run
repo forall -c git log --oneline (or similar) and see what happened recently for each project separately.
There also seems to be a subcommand that gives a changelog for all projects together. I haven’t tried it yet.
In any case, marking a particular point in time before
syncing is possible with tags. For instance:
$ repo forall -c git tag REMEMBER_THIS_COMMIT
Another way of watching new commits is by partially syncing with
repo sync -n. That fetches all commits but leaves
HEAD (your working directory) at the current state.
These are but building blocks and I have not found a satisfactory way of getting a definite changelog.
Thanks. I will sync -n and play around with the repo extension. I think by now git can also handle multiple git repos itself.
Anyway, there has to be some working branches/tags or trunks to get code for the current image. It cannot be that hard. All the fixed CVEs should show up in there.
AAaaaaaand another week passed by without any information… Or are there any news about the fposos??
Just have a look here: Fairphone Open Source OS is almost ready. Want to test?
Perhaps this is said before. I would love to use Fairphone Open Source if it is easy to root and there is a easy to understand explanation how to get some Google apps to work on the Fairphone Open Source if there’ no alternative for the app in F-droid and others (like the NS reisplanner, Dutch traintables).
If that’s the case I guess people will more willing to give it a try.
The problem is the company that builds the “NS reisplanner”. They use the Google framework for messaging and for apps.
I’m not sure if there are good alternatives for it (Open Street Map for maps, but as far as I know nothing for messaging), but in the meantime you could use microG. But as long as companies will build services that depend on the Google framework, users will depend on the Google services.
Yes, I know the company NS is the problem.
I am wondering if something like Evozi’s APK Downloader is an alternative, considering malware etc.?
There’s an app in F-Droid which may cover your usecase: Transportr
For the Google-jail problem, there is BlankStore, part of the NoGapps project (which evolved into microG project, microg.org).
But easiest is flashing the smallest of the Google Apps package (OpenGapps pico looks small enought).
Your choose: freedom and privacy but complexity versus dependence and not-so-much-privacy but easiness
This topic indeed has been discussed multiple times but there are not easy solutions for your use case. By installing even a minimal set of google tools you give away the freedom you got by installing the open source version.
Ideally the thing to do is to convince NS and such that they need to offer their services for all users and not only users with a google account. You then also need to convince them to not use google specific interfaces and we need to offer a way to keep things up-2-date for the user.
If this is not possible I can think of two alternatives:
-Use multiple accounts on a single phone (one that is a google account and one account that know nothing about the google login). It would be very nice if it where possible to disable gapps on the second account
-Use a second phone (or tablet) where the apps are being downloaded and copy over the APK files: this is what I currently do: I use apk backup to store my apks on a server (seafile) and install them on my “free phone” (arguably not everybody has access to a second phone).
What is really lacking is a trustworthy (f-droid style) app store for commercial apps
Update: I rewrote this because it looks like it caused misunderstandings @keesj.
[Using apps like NS reisplanner that use the Google framework … ‘without’ Google]
There is no “without” Google for these apps (your phone still talks to the Google servers), but you can make these apps work and also remove the Google services on your phone. Technical, using mircoG and Raccoon should also work. I tried it and it looked okay. Only the real time map is pretty slow.
You can use a lot of apps on your phone without Google. But apps the were build with the Google services framework and that do not offer a fallback, still need the Google servers to function.
@keesj and @fp1_wo_sw_updates thank you both. As I have Linux (Fedoria) on my computer I love to have Fairphone Open Source OS on my Fairphone.
Hopefully the tests are succesfull and the release is soon there.
I have used many tools on my phone without having a google account, this include apps like ov9292 or similar. Stating there is no ‘without’ google is not true.
Can you elaborate what you mean by that? Which parts of Android talk to google services when none of the google services have been installed?
The only thing I can think of is the captive portal check, which IIRC uses a google server to check internet connectivity.
I was talking about the software @Lidwien is using: NS reisplanner
This app is using the Google framework for maps and messages.
Phones without Google services or apps that do not depend on Google services are possible.
I’ve never said that. Or lets say I did not said it very well. I was still talking about apps the are using the Google framework. This is what makes it hard for the users.
I’m sorry for the confusion I caused. In this case I was mainly talking about apps the use the google services framework for maps and messaging. In this case particular I was talking about apps like NS reisplanner. Apps like this one talk to Google services to be able to fully work (They look up were you are, embedded Google maps and uses the Google messaging service). But the programmers could make them also work without Google. But often they do not offer a fallback mode.
This makes it hard for the users. They only “see” that their app is not fully working.