| 7 Sep 2021 |
 kranzes | 
Download image.png | 20:25:28 |
| kranzes | but this is what's left | 20:25:33 |
| kranzes | In reply to @kranzes:matrix.org
i highly doubt that does anything in this case* | 20:25:43 |
| kranzes | forklineageos uses a different android_vendor_lineage: https://github.com/ForkLineageOS/android_vendor_lineage | 20:27:20 |
| kranzes | What do you think we should do next? | 20:29:11 |
 danielrf | Two things we could do next: 1) dig through forklineageos changes that could be affected by either RELEASE_TYPE or OVERRIDE_TARGET_FLATTEN_APEX, or 2) continue bisecting by uncommenting RELEASE_TYPE and rebuilding/retesting | 20:31:14 |
| danielrf | Actually one other way we could try: compare the extracted images between working and non-working robotnix builds | 20:34:12 |
| kranzes | every mention of OVERRIDE_TARGET_FLATTEN_APEX in ForkLineageOS is exactly the same as upstream LineageOS | 20:34:23 |
| kranzes | In reply to @danielrf:matrix.org
Actually one other way we could try: compare the extracted images between working and non-working robotnix builds Not sure how to do that | 20:34:36 |
| danielrf | Now that it's building in robotnix, I'd be more confident that our comparisons could reveal something | 20:34:40 |
| kranzes | Do you want me to upload a working robotnix built image? | 20:35:11 |
| kranzes | and a a working docker build image? | 20:35:19 |
| danielrf | It's OK. I'll try building both myself | 20:35:35 |
| kranzes | In reply to @danielrf:matrix.org
It's OK. I'll try building both myself Life saving | 20:35:46 |
| kranzes | if i was to build now again, | 20:35:52 |
| kranzes | it will only be done at 3 AM | 20:35:57 |
| kranzes | 6-7 AM if im gonna try building two images | 20:36:28 |
| kranzes | for you it will take like what, 1 Hour? for both | 20:36:49 |
| danielrf | eh, maybe closer to 1.5hrs, we'll see | 20:37:06 |
| kranzes | does signing.apex.enable = false; depend on setting envVars.OVERRIDE_TARGET_FLATTEN_APEX = "true"; ? | 20:37:34 |
| danielrf | They should be enabled/disabled together | 20:38:07 |
| kranzes | 👍️ | 20:38:15 |
| danielrf | kranzes: Can you verify if the following image works? https://drive.google.com/file/d/1sumK24CzbWitR2XKSxUDatvYMyrkdVYP/view?usp=sharing | 21:25:40 |
| danielrf | Here's the diff between those two images, one with the APEX flattening / RELEASE_TYPE commented out, and one without | 21:28:25 |
| danielrf | it's basically what you'd expect. SELinux changes for APEX, ro.lineage.releasetype differences, etc. | 21:29:01 |
| danielrf | One thing that stands out a little is that some libraries are only in one image vs. the other. e.g. libcrypto_utils.so, libincident.so | 21:29:55 |
| danielrf | and I wonder if that's a result of enabling/disabling APEX flattening | 21:30:09 |
| kranzes | In reply to @danielrf:matrix.org
kranzes: Can you verify if the following image works? https://drive.google.com/file/d/1sumK24CzbWitR2XKSxUDatvYMyrkdVYP/view?usp=sharing what changes does this image have? | 21:33:24 |
| danielrf | That image just has commented out APEX flattening / RELEASE_TYPE. | 21:33:50 |
| kranzes | didnt we build that before then? | 21:34:13 |
