summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* android-x86: override libaudioreplicant-10-devDenis 'GNUtoo' Carikli2020-11-191-1/+1
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* android-x86: fix revision typoDenis 'GNUtoo' Carikli2020-11-181-1/+1
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* android-x86: make the provenance more clearDenis 'GNUtoo' Carikli2020-11-181-1/+1
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* android-x86: fix base repo linkDenis 'GNUtoo' Carikli2020-11-181-1/+1
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Add android-x86 Android 10 AudioDenis 'GNUtoo' Carikli2020-11-183-0/+15
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Add libsamsung-rilDenis 'GNUtoo' Carikli2020-10-151-0/+1
| | | | | | | | | | | | | libsamsung-ril should now properly compile under Replicant 10, however the patches still need to be tested at runtime under Replicant 10, so for the time being, libsamsung-ril is not yet built by default for the GT-I9300. The libsamsung-ril patches also need to be tested on Replicant 6.0, and Replicant 4.2 to make sure they still build and don't break anything. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Add external/evtestDenis 'GNUtoo' Carikli2020-10-121-0/+1
| | | | | | | | | | | This can help debug software and hardware issues so it's best to ship it as it is time consuming to build an image just for adding it. Evtest has to be added to PRODUCT_PACKAGES to be shipped on the devices. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Override build/soonDenis 'GNUtoo' Carikli2020-09-042-5/+4
| | | | | | | | | | | | | | | | | | | | | | | | | This is done for two reasons: - Building mesa even with TEMPORARY_DISABLE_PATH_RESTRICTIONS=true results in a huge quantity of warnings like that: "xgettext" is not allowed to be used. See https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools for more information. - Not enabling TEMPORARY_DISABLE_PATH_RESTRICTIONS might even work if we manage to whitelist all the tools we need. It would also be a good thing as we need to use the host python2 and python3 to build mesa, and many other scripts in the Android source simply use 'python'. This will make sure (without having to dig into the details) that the host 'python' is not used accidentally. It could be an issue if the host 'python' points to python3, as Android 10 expects python to be python2. For instance: $ python3 ./bionic/libc/tools/genseccomp.py Traceback (most recent call last): File "./bionic/libc/tools/genseccomp.py", line 11, in <module> from gensyscalls import SysCallsTxtParser File "/home/replicant/replicant-10/bionic/libc/tools/gensyscalls.py", line 410 print "%d: %s" % (self.lineno, msg) ^ Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* vendor/lineage: link scripts to vendor/replicantDenis 'GNUtoo' Carikli2020-08-281-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Replicant has decided to not move vendor/lineage to vendor/replicant. It's not possible to symlink vendor/lineage to vendor/replicant as some files would end up being included twice. The short terms benefits for moving it was that it would be consistent with the previous Replicant versions and that the build scripts that are already used in the build tutorials would be in a place where users do expect them. However the long terms benefits of not moving it are way bigger: * It lowers the amount of maintenance and the amount of work for keeping the current version up to date and to make new Replicant version in a very significant way: * Moving it requires to fork many repository just for changing the path to vendor/replicant * All the repositories will need to be rebased or merged which is time consuming * Rebasing vendor/replicant itself is usually a pain because of the numerous changes just to change vendor/lineage to vendor/replicant but also because of some variables names changes from LINEAGE/LINEAGEOS to REPLICANT. So symlinking just the scripts should be enough to make every use case work. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Use Replicant mirrors for AOSP repositoriesDenis 'GNUtoo' Carikli2020-08-281-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We need to mirror AOSP repositories to make sure that we don't loose the source code. The git hosting infrastructure has not been adjusted yet to automatically redirect repositories like /device/generic/qemu to device_generic_qemu in /mirrors/AOSP/. The redirect is manual for now, which means that for each new repository that you add you must not forget to do the redirection in /etc/apache2/git-redirects/aosp-mirror.conf else it will not work. To make an automatic redirection we need to change the '/' in a '_', but also find where to stop the match. The URL itself doesn't have enough information to be able to do that correctly. With a simple rule like 'RewriteRule "^/mirrors/AOSP/(.*)/(.+)" "/mirrors/AOSP/$1_$2" [R]' device/generic/qemu/tree/ would be rewritten as device_generic_qemu_tree instead of device_generic_qemu/tree and using heuristics to workaround that is probably not a good idea as it would be way to fragile. Currently, the static list of redirect has been generated with the following command: ls -d *.git | sed 's#\.git$##' | \ awk '{save=$1 ; gsub("_", "/", $1); printf("Redirect /mirrors/AOSP/%s /mirrors/AOSP/%s\n", $1, save); }' So it should be possible to generate to convert that to a single set of regular expressions that relies on checking if specific directories do exist or not. However this is not that simple as we want to avoid potential bugs that only appear in corner case and that could take a lot of time to debug. For instance if the device_generic_qemu.git directory is matched in the file system, the regex should stop converting '/' in '_', however the regex needs to start to match from the right as we don't want to introduce bugs in the unlikely case where we have something like the following repositories (instead of eclipse-build-deps): - platform_prebuilts_eclipse - platform_prebuilts_eclipse_build_deps Another way would be to use inotify/fanotify scripts to generate the mapping automatically. Having something that works would enable people not to have to learn the details on how this system work, especially if they are not aware it exists and have to learn that during days of debugging without even having ssh access. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* libsamsung-ipc: clarify that it comes from the main Replicant namespaceDenis 'GNUtoo' Carikli2020-08-282-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | As libsamsung-ipc master is now supposed to compile under Replicant 10, it has been included and enabled for the i9300 target to be built tested. The focus is now to merge more and more patches upstream in the master branch before adding an abstraction for the modem driver in the Replicant 9 or 10 kernel (which may or may not land in a branch). Because of that, the replicant-next repository of libsamsung-ipc contains outdated code, and the most recent code is in libsamsung-ipc master. It also didn't compile with the replicant-next repository as all the fixes for that weren't merged. Because of all that, we are now using libsamsung-ipc from the master of the main replicant namespace. However, having remote="replicant-next" is confusing and prone to mistakes. While it selects the exact same repository thanks to name="replicant/...", at some point in the history that has been saved in the history/28-08-2020 branch, it was accidentally changed to use the replicant-next namespace probably with sed by the following commit: 7747d021e3fd8c7b93b8aa9f25199130018e4c00 7747d021 move vendor/lineage to vendor/replicant Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Checked out manifest: Use Replicant's manifest and adjust to release scriptsDenis 'GNUtoo' Carikli2020-08-282-1/+1
| | | | | | | The Replicant release scripts, more specifically like the tagging script rely having the manifest in manifest/. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* replicant: set default revision to replicant-10Denis 'GNUtoo' Carikli2020-08-282-12/+12
| | | | | | | | | | | This makes the replicant repositories specifications shorter and less error prone as you don't need not to forget to set a revision anymore unless you intend to override the default revision. It also makes it more obvious when repositories have a non default revision as it stands more out. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* replicant: repositories: hardware_samsung: fix namingDenis 'GNUtoo' Carikli2020-08-281-1/+1
| | | | | | | | | By convention, in replicant/ or replicant-next/, the repositories are named after the path, while in LineageOS they are often prefixed with android_. The repository also has been moved in the git hosting infrastructure. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* replicant: order repositories aphabetically by pathDenis 'GNUtoo' Carikli2020-08-281-5/+5
| | | | Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Update drm_hwcomposer and gbm_gralloc versionsDavid Ludovino2020-08-281-2/+2
| | | | | Signed-off-by: David Ludovino <dllud@riseup.net> Change-Id: I5a367e68a93265bee4e92402cb941238256cd1df
* Handle fetch from contrib/replicant-next/ instead of replicant/Denis 'GNUtoo' Carikli2020-08-281-2/+2
| | | | | | | | | | | | | | | | | | | | Using '..' enables repo to use URL relative to the location it was fetched from. This is really handy as we can enable multiple protocol with the same git repository: - http:// is useful for potential caching on a network - https:// is probably more secure and privacy friendly and is typically not blocked by restrictive network policies. - git:// is probably secure, privacy friendly and more efficient and sometimes blocked by restrictive network policies. However the side effect is that it makes the fetch relative to the URL it is cloned from, which can create issues if the depth level is changed. This works around such issue for repositories not yet in the main namespace. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Move Replicant specific repos to new path: replicant-nextDavid Ludovino2020-08-281-12/+12
| | | | | Signed-off-by: David Ludovino <dllud@riseup.net> Change-Id: I912984ce2c68c2994b1b0fac9ac4f056bb9378b1
* Use new location for Replicant 10 development repositoriesJoonas Kylmälä2020-08-281-2/+2
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I41f364cac4a14b62c3e059f7969e5a1c39b4c0f9
* Update drm_hwcomposer and gbm_gralloc versionsJoonas Kylmälä2020-08-281-2/+2
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: Ia0bc3c495a2b2cd1535a33494b5c273964e021c2
* Use Replicant's fork of hardware/samsungJoonas Kylmälä2020-06-252-2/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I3fbf81d5c6a170c3b498daa5ab0bc1de9b157492
* Use mirrors for LineageOS repositoriesDenis 'GNUtoo' Carikli2020-06-253-7/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | LineageOS repositories don't use stable revisions and because of that, the changes we have on top of LineageOS repositories need to be rebased often as they need to be kept in sync with the other repositories. While it might be a good thing to constantly rebase in order to get the last security fixes, this is not relevant at this early stages of development where the devices are not used daily. Instead this makes us loose time and makes it harder to contribute. The other repositories could be mirrored too however they all use stable revisions: - All the repositories using the github and freedesktop remotes use fixed revisions. - All the AOSP repositories use a tag The only exception being tools/repohook which is set to use the master revision. The git hosting infrastructure has also been adjusted to enable to use git://git.replicant.us/mirrors/LineageOS/* and https://git.replicant.us/mirrors/LineageOS/* in order to make the manifest patch smaller and easier to rebase. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* Use Replicant's fork of libdrmJoonas Kylmälä2020-06-252-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I6925c03a862f2a9f9f43156c58d47eef6c0c5106
* Include midas-common repositoryJoonas Kylmälä2020-06-251-0/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I1081291cde04e70867020e1e0addd253e5aede79
* Add i9300 device treeJoonas Kylmälä2020-06-251-0/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I8025fd1448db558f50abf19d9e6fa493af965a2d
* Use Replicant 10 fork of build/makeJoonas Kylmälä2020-06-251-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: Ib588b64ca2e83b484c1b0e03e0688ceebde1be43
* Use Replicant 10 fork of kernel and mesa3dJoonas Kylmälä2020-06-251-2/+2
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I190425a2bbd87a7d135a0877a3482f2671a459c1
* Use Replicant 10 branch of device_samsung_i9305 repoJoonas Kylmälä2020-06-251-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I272953e93277004e01580306ac5fe958218daf33
* Use Replicant 10 branches for our forked repositoriesJoonas Kylmälä2020-06-251-3/+3
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: If1c6c6bc9aa7f2825dafdec0605cef5f0e3afe9b
* Use AOSP version of libdrmJoonas Kylmälä2020-06-251-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I38bb14592e4608e752108fff51dde3fc138fc379
* Use latest master version of SwiftShaderJoonas Kylmälä2020-06-251-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I18f6189cc1dad2cd9828dbfdb432cac2557fd22f
* Correct libsamsung-ipc path after it was renamedJoonas Kylmälä2020-06-251-1/+1
| | | | | Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi> Change-Id: I91d2a0003db542ce9b9c51656f52f67bcef7d889
* Add libsamsung-ipc for testingDenis 'GNUtoo' Carikli2020-06-251-0/+1
| | | | | | | | | | | | | | | | | | | | | At the time of writing, libsamsung-ipc is still not ready for Replicant 9 as there is still ongoing work to cleanup and integrate the patches that makes the modem work with the replicant 9 kernel driver. However since this integration work is going on, we also need to make sure that libsamsung-ipc also compiles on Replicant 9. This change will ensure that libsamsung-ipc is built and shipped inside the image, so if the compilation breaks at some point, it will enable to detect it more easily. The kernel driver and libsamsung-ril also need to be cleaned up, and the ability to launch libsamsung-ipc twice without crashing the device needs to be implemented. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* SwiftShader revision that uses LLVM instead of SubZero.David Ludovino2020-06-251-1/+1
| | | | | | | | | | | This is the latest SwiftShader revision that still works with the frameworks/native on AOSP 9. Having LLVM as the single backend JIT compiler facilitates our task of finding out why SwiftShader is emitting UDIV/SDIV instructions on hardware that does not support them. Change-Id: Ic08668bbedc5ea79479598510c250f0413fc9533 Signed-off-by: David Ludovino <dllud@riseup.net>
* Add hardware/samsungJoonas Kylmälä2020-06-251-0/+3
| | | | | | | It provides useful HAL modules like liblights. Change-Id: I2594e377f8fee1d695e95f33864cebb44971dc81 Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi>
* Moved frameworks_base into Replicant maintained repos.David Ludovino2020-06-252-1/+1
| | | | | | | | This allows us to patch it in order to better support software rendering. Change-Id: I080fff3001d73e0e07acecfd91cdddeba6045774 Signed-off-by: David Ludovino <dllud@riseup.net>
* Add Replicant repositoriesDenis 'GNUtoo' Carikli2020-06-253-14/+32
| | | | | | | | | | | | | | At first we tried to move LineageOS's default.xml in a subdirectory to keep as little changes as possible with LineageOS in that file. However git didn't properly detect that the file was moved as some changes in the files were necessary. As the Replicant includes are now at the end of the file, it should minimize the non-automatic rebases. Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org> Rebased the commit and adapted it for LineageOS 17.1 Signed-off-by: Joonas Kylmälä <joonas.kylmala@iki.fi>
* Fix fetch from githubDenis 'GNUtoo' Carikli2020-06-251-1/+1
| | | | | | | | | | | | | | | | | | | | The fetch=".." directive is relative to the URL repository hosting the manifest, as without this fix, a repository like android_build would be fetched from https://git.replicant.us/LineageOS/android_build.git. The manifest.xml doesn't have any reference to git.replicant.us. The only reference to git.replicant.us is from the URL of the manifest repository which in this case is is at the following URL: https://git.replicant.us/GNUtoo/manifest.git. This behavior is however not documented in the repo manifest format[1]. References: ----------- [1]https://android.googlesource.com/tools/repo/+/refs/tags/v1.13.2/docs/repo-hooks.md Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
* manifest: Track additional projects introduced in android-10.0.0_r37Kevin F. Haggerty2020-06-021-0/+1
| | | | Change-Id: I8d091d335f88c76048639307f592dcbe42bb7cf0
* manifest: android-10.0.0_r35 -> android-10.0.0_r37Kevin F. Haggerty2020-06-021-1/+1
| | | | Change-Id: Ic805bbef0d77dcffb21785a6c9354da1cb896300
* manifest: Switch to our external/libtextclassifier forkHan Wang2020-06-011-1/+1
| | | | Change-Id: Iec83239ce5f47e8625c0d6dea6fc001c483ebd7c
* android: Move to our own fork of timezonepickerMichael W2020-05-241-1/+1
| | | | Change-Id: I699809d128d21cbc48ef32730baac57d43a54f97
* Remove unused frameworks/opt/datetimepickerNeil Fuller2020-05-241-1/+0
| | | | | | | | | | The code has Y2038 issues via android.text.format.Time. There are no users so it can be removed. Bug: 16550209 Test: build only Merged-In: Ic07eb1f810d5192d69b156fff6d64ff538471651 Change-Id: Ic07eb1f810d5192d69b156fff6d64ff538471651
* manifest: Replace Calendar with EtarLuK13372020-05-242-1/+1
| | | | Change-Id: If8a5786ee40cc8a20931eddfee51ed028155315a
* Fetch our fork of build-toolsChirayu Desai2020-05-141-1/+1
| | | | Change-Id: I2191ba7d798fffb5bf5792de3cdb9034b144fd0f
* manifest: Use our forks of e/{cldr,icu} and s/timezoneKevin F. Haggerty2020-05-101-3/+3
| | | | | | * Necessary for tzdb updates Change-Id: I19249061a6bfd573af9054139ddb507dbddb6876
* manifest: Move from android-10.0.0_r33 to android-10.0.0_r35Davide Garberi2020-05-051-1/+1
| | | | Change-Id: I8c5d4f2900e138bf5b1d87426e466d9f0af0aa74
* manifest: Sync vendor/qcom/opensource/vibratorLuK13372020-05-051-0/+1
| | | | Change-Id: I31a4ec3f4838314d5bb6c4854c2e5aa6a825c0a1
* manifest: Track additional Pixel 4{,XL} hardware repoNolen Johnson2020-05-011-1/+1
| | | | Change-Id: I2b2b4ddef7ab08fabf8c2fc1164084251c94b24c
* manifest: Track packages/inputmethods/LeanbackIMEAaron Kling2020-04-231-0/+1
| | | | Change-Id: I346510648da55a50ebc03fc4150317715d7e92f2