[android-building] Missing dep libraries in generated apex

2019-10-03 Thread Michael Goffioul 
Hi,

I'm trying to build Android Q (for android-x86, 32-bits) and I'm having a 
problem in the generated apex for com.android.media.swcodec, where there 
are a bunch of libraries missing in 
$TARGET_OUT/system/apex/com.android.media.swcodec/lib/.

The directory does contain various deps like libcodec2_h...@1.0.so, 
libmedia_codecserviceregistrant.so or libstagefright_amrnb_common.so. 
However it is missing some, most 
notably libstagefright_bufferqueue_helper.so or all libcodec2_soft_*.so 
modules. This then results in runtime errors when running mediaswcodec 
service.

What can be the reasons for deps to be missing from the apex?

Note that I tried to find my way in build/soong/, but didn't get anywhere 
in trying to debug this issue.

Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/9c01756b-cc89-4968-95dd-de1563efc801%40googlegroups.com.

[android-building] Re: Missing dep libraries in generated apex

2019-10-04 Thread Michael Goffioul 
Answering myself...
It turns out the reason the deps are not included in the APEX is the 
presence of cfi sanitizer in their Android.bp. When this is set, 
HideFromMake for the module is set to true, and the module is then not 
included when generating out/soong/Android-${TARGET_PRODUCT}.mk

On Thursday, October 3, 2019 at 3:53:13 PM UTC-4, Michael Goffioul wrote:
>
> Hi,
>
> I'm trying to build Android Q (for android-x86, 32-bits) and I'm having a 
> problem in the generated apex for com.android.media.swcodec, where there 
> are a bunch of libraries missing in 
> $TARGET_OUT/system/apex/com.android.media.swcodec/lib/.
>
> The directory does contain various deps like libcodec2_h...@1.0.so, 
> libmedia_codecserviceregistrant.so or libstagefright_amrnb_common.so. 
> However it is missing some, most 
> notably libstagefright_bufferqueue_helper.so or all libcodec2_soft_*.so 
> modules. This then results in runtime errors when running mediaswcodec 
> service.
>
> What can be the reasons for deps to be missing from the apex?
>
> Note that I tried to find my way in build/soong/, but didn't get anywhere 
> in trying to debug this issue.
>
> Michael.
>
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/b114854b-ef0a-461e-b37f-8cc9f98c20ce%40googlegroups.com.

[android-building] Unable to download content during build

2019-10-15 Thread Michael Goffioul 
I have several custom modules that download content (if not already 
present) during build, with wget. This uses additional rules in the 
corresponding Android.mk files. These used to work in Pie, but in Q the 
rule fail with "Name or service not found". Running the exact same command 
outside the build system runs fine. Can somebody describe the mechanism at 
play here, and maybe points to source files where the restriction is 
implemented?

Thanks,
Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/403e25aa-3db5-499a-a7f7-e05814b66fa6%40googlegroups.com.

Re: [android-building] Unable to download content during build

2019-10-16 Thread Michael Goffioul 
Thanks, I figured it must be somehow block. What I would like to understand
is _how_ this is achieved technically.


On Wed, Oct 16, 2019 at 12:58 PM 'Wesolowski, Krzysztof' via Android
Building  wrote:

> It's blocked by sandboxing of build system, isolating that only repo
> content is used. Probably wget or external connections are blocked.
>
> BR, K
>
> Get Outlook for Android <https://aka.ms/ghei36>
>
> --
> *From:* android-building@googlegroups.com <
> android-building@googlegroups.com> on behalf of Michael Goffioul <
> michael.goffi...@gmail.com>
> *Sent:* Tuesday, October 15, 2019 9:58:18 PM
> *To:* Android Building 
> *Subject:* [android-building] Unable to download content during build
>
> I have several custom modules that download content (if not already
> present) during build, with wget. This uses additional rules in the
> corresponding Android.mk files. These used to work in Pie, but in Q the
> rule fail with "Name or service not found". Running the exact same command
> outside the build system runs fine. Can somebody describe the mechanism at
> play here, and maybe points to source files where the restriction is
> implemented?
>
> Thanks,
> Michael.
>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
> <https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgroups.google.com%2Fgroup%2Fandroid-building%3Fhl%3Den&data=02%7C01%7Ckrzysztof.wesolowski%40volvocars.com%7C04756d551b9d41282daa08d752014292%7C81fa766ea34948678bf4ab35e250a08f%7C0%7C0%7C637068037674699522&sdata=Dn3KNRQFJustHPoSAd3cceR%2Bp%2BlBzdedzBRB4VDwJ44%3D&reserved=0>
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/403e25aa-3db5-499a-a7f7-e05814b66fa6%40googlegroups.com
> <https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fandroid-building%2F403e25aa-3db5-499a-a7f7-e05814b66fa6%2540googlegroups.com%3Futm_medium%3Demail%26utm_source%3Dfooter&data=02%7C01%7Ckrzysztof.wesolowski%40volvocars.com%7C04756d551b9d41282daa08d752014292%7C81fa766ea34948678bf4ab35e250a08f%7C0%7C0%7C637068037674699522&sdata=xNA99VclgJ5ZT%2B0H2Y1AzwvgG9voZJrpq3rULPMbajc%3D&reserved=0>
> .
>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/VE1P174MB00787459500593F9D0B659AC91920%40VE1P174MB0078.EURP174.PROD.OUTLOOK.COM
> <https://groups.google.com/d/msgid/android-building/VE1P174MB00787459500593F9D0B659AC91920%40VE1P174MB0078.EURP174.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer>
> .
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LsvY8ffMGZhOTmGEn_gmyd8sGnXyhtX84zoqAxy%2B93DFw%40mail.gmail.com.

[android-building] Lots of 'Disallowed PATH tool "manpath"...' when building Android Q

2019-11-14 Thread Michael Goffioul 
I'm using Fedora 30 to build AOSP. Since Android Q, when building, the 
terminal output is littered with hundreds of messages like the following:

17:23:06 Disallowed PATH tool "manpath" used: []string{"manpath"}
17:23:06 See 
https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools 
for more information.

I've looked almost everywhere, but I can't find where "manpath" is actually 
used. Has anybody encountered the same problem?

Thanks,
Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/c9a8a0ea-2492-4e7f-9816-11b9b91f6696%40googlegroups.com.

Re: [android-building] Lots of 'Disallowed PATH tool "manpath"...' when building Android Q

2019-11-18 Thread Michael Goffioul 
Yes, I've already looked into that, but there's not much of a trace, e.g.:

2019/11/18 11:22:34.837088 build/soong/ui/build/sandbox_linux.go:86:
[prebuilts/build-tools/linux-x86/bin/nsjail -H android-build -e -u nobody
-g nobody -B / --disable_clone_newcgroup -- /bin/bash -c if [ $(hostname)
== "android-build" ]; then echo "Android" "Success"; else echo Failure; fi]
2019/11/18 11:22:34.877084 build/soong/ui/build/path.go:105: Disallowed
PATH tool "manpath" used: []string{"manpath"}
2019/11/18 11:22:34.877237 build/soong/ui/build/path.go:107: See
https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
for more information.

I've a feeling that it's coming from bash itself. I've looked into the
local and system profile/bashrc and anything I could think of, but I can't
really pinpoint where it's coming from.

On Mon, Nov 18, 2019 at 12:40 PM 'Dan Willemsen' via Android Building <
android-building@googlegroups.com> wrote:

> If you check out/soong.log, there should be trace logs that include the
> parent processes all the way up to ninja.
>
> It sounds like one of the existing tools that we allow through is always
> trying to call manpath? Let me know what you find, we've definitely been
> shrinking this list on master.
>
> - Dan
>
> On Thu, Nov 14, 2019 at 3:54 PM Michael Goffioul <
> michael.goffi...@gmail.com> wrote:
>
>> I'm using Fedora 30 to build AOSP. Since Android Q, when building, the
>> terminal output is littered with hundreds of messages like the following:
>>
>> 17:23:06 Disallowed PATH tool "manpath" used: []string{"manpath"}
>> 17:23:06 See
>> https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
>> for more information.
>>
>> I've looked almost everywhere, but I can't find where "manpath" is
>> actually used. Has anybody encountered the same problem?
>>
>> Thanks,
>> Michael.
>>
>> --
>> --
>> You received this message because you are subscribed to the "Android
>> Building" mailing list.
>> To post to this group, send email to android-building@googlegroups.com
>> To unsubscribe from this group, send email to
>> android-building+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/android-building?hl=en
>>
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "Android Building" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to android-building+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/android-building/c9a8a0ea-2492-4e7f-9816-11b9b91f6696%40googlegroups.com
>> <https://groups.google.com/d/msgid/android-building/c9a8a0ea-2492-4e7f-9816-11b9b91f6696%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/CALQgHd%3DmqvANYjYSZyP_1%3DtbMgoddYSLEMTkn3NuUTJF_qML7A%40mail.gmail.com
> <https://groups.google.com/d/msgid/android-building/CALQgHd%3DmqvANYjYSZyP_1%3DtbMgoddYSLEMTkn3NuUTJF_qML7A%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LtJOhcEKr8m_nExyByvOzb_opZxJvk4xovia7d-oYkhdA%40mail.gmail.com.

Re: [android-building] Lots of 'Disallowed PATH tool "manpath"...' when building Android Q

2019-11-18 Thread Michael Goffioul 
Thanks, that worked :)


On Mon, Nov 18, 2019 at 2:16 PM 'Dan Willemsen' via Android Building <
android-building@googlegroups.com> wrote:

> I've created a change to empty out the BASH_ENV environment variable
> during the build:
> https://android-review.googlesource.com/c/platform/build/soong/+/1168864
>
> - Dan
>
> On Mon, Nov 18, 2019 at 10:48 AM 'Colin Cross' via Android Building <
> android-building@googlegroups.com> wrote:
>
>> I was able to reproduce this on Fedora 30, it's something to do with the
>> environment BASH_ENV=/usr/share/Modules/init/bash, which is causing
>> arbitrary bash code that can use any binary to be run on every
>> non-interactive bash shell startup.
>>
>> On Mon, Nov 18, 2019 at 9:50 AM Michael Goffioul <
>> michael.goffi...@gmail.com> wrote:
>>
>>> Yes, I've already looked into that, but there's not much of a trace,
>>> e.g.:
>>>
>>> 2019/11/18 11:22:34.837088 build/soong/ui/build/sandbox_linux.go:86:
>>> [prebuilts/build-tools/linux-x86/bin/nsjail -H android-build -e -u nobody
>>> -g nobody -B / --disable_clone_newcgroup -- /bin/bash -c if [ $(hostname)
>>> == "android-build" ]; then echo "Android" "Success"; else echo Failure; fi]
>>> 2019/11/18 11:22:34.877084 build/soong/ui/build/path.go:105: Disallowed
>>> PATH tool "manpath" used: []string{"manpath"}
>>> 2019/11/18 11:22:34.877237 build/soong/ui/build/path.go:107: See
>>> https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
>>> for more information.
>>>
>>> I've a feeling that it's coming from bash itself. I've looked into the
>>> local and system profile/bashrc and anything I could think of, but I can't
>>> really pinpoint where it's coming from.
>>>
>>> On Mon, Nov 18, 2019 at 12:40 PM 'Dan Willemsen' via Android Building <
>>> android-building@googlegroups.com> wrote:
>>>
>>>> If you check out/soong.log, there should be trace logs that include the
>>>> parent processes all the way up to ninja.
>>>>
>>>> It sounds like one of the existing tools that we allow through is
>>>> always trying to call manpath? Let me know what you find, we've definitely
>>>> been shrinking this list on master.
>>>>
>>>> - Dan
>>>>
>>>> On Thu, Nov 14, 2019 at 3:54 PM Michael Goffioul <
>>>> michael.goffi...@gmail.com> wrote:
>>>>
>>>>> I'm using Fedora 30 to build AOSP. Since Android Q, when building, the
>>>>> terminal output is littered with hundreds of messages like the following:
>>>>>
>>>>> 17:23:06 Disallowed PATH tool "manpath" used: []string{"manpath"}
>>>>> 17:23:06 See
>>>>> https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
>>>>> for more information.
>>>>>
>>>>> I've looked almost everywhere, but I can't find where "manpath" is
>>>>> actually used. Has anybody encountered the same problem?
>>>>>
>>>>> Thanks,
>>>>> Michael.
>>>>>
>>>>> --
>>>>> --
>>>>> You received this message because you are subscribed to the "Android
>>>>> Building" mailing list.
>>>>> To post to this group, send email to android-building@googlegroups.com
>>>>> To unsubscribe from this group, send email to
>>>>> android-building+unsubscr...@googlegroups.com
>>>>> For more options, visit this group at
>>>>> http://groups.google.com/group/android-building?hl=en
>>>>>
>>>>> ---
>>>>> You received this message because you are subscribed to the Google
>>>>> Groups "Android Building" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>>> an email to android-building+unsubscr...@googlegroups.com.
>>>>> To view this discussion on the web visit
>>>>> https://groups.google.com/d/msgid/android-building/c9a8a0ea-2492-4e7f-9816-11b9b91f6696%40googlegroups.com
>>>>> <https://groups.google.com/d/msgid/android-building/c9a8a0ea-2492-4e7f-9816-11b9b91f6696%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>> --
>>>>

[android-building] Disallowed PATH tool "m4", but used by flex

2020-03-30 Thread Michael Goffioul 
I'm experimenting with android-r-preview-2 and found a problem compiling
mesa, which didn't occur with android-10. The error occurs when running
flex:

23:56:24 Disallowed PATH tool "m4" used: []string{"m4", "-P"}
23:56:24 See
https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
for more information.
[  0% 1/74788] Mesa Lex: libmesa_glsl <=
external/mesa/src/compiler/glsl/glcpp/glcpp-lex.l
FAILED:
out/target/product/x86/gen/STATIC_LIBRARIES/libmesa_glsl_intermediates/glsl/glcpp/glcpp-lex.c
/bin/bash -c "prebuilts/build-tools/linux-x86/bin/flex --nounistd
-oout/target/product/x86/gen/STATIC_LIBRARIES/libmesa_glsl_intermediates/glsl/glcpp/glcpp-lex.c
external
/mesa/src/compiler/glsl/glcpp/glcpp-lex.l"
"m4" is not allowed to be used. See
https://android.googlesource.com/platform/build/+/master/Changes.md#PATH_Tools
for more information.

>From what I understand, flex internally calls "m4" (according to this
reference: https://www.cs.virginia.edu/~cr4bd/flex-manual/M4-Dependency.html
).

Any pointer in the right direction would be appreciated, thanks.

Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LtxMUAWGZG2SbnjKB0iOiNAQV5R5oZP%2BHaUaeeW%3D1LZMA%40mail.gmail.com.

[android-building] Re: Dex2oat failed to compile a boot image

2020-03-30 Thread Michael Goffioul 
I found out I was missing this change:
https://android.googlesource.com/platform/libcore/+/271b4054b825efca3176c4ac11565e33cff5d09a
I cherry-picked it, but now I'm getting a different error:

dex2oatd F 03-30 19:11:33  6962  6962 unstarted_runtime.cc:2085] Calling
native method void android.system.OsConstants.initConstants() in an
unstarted non-transactional runtime


On Mon, Mar 30, 2020 at 6:59 PM Michael Goffioul 
wrote:

> On Mon, Mar 30, 2020 at 6:52 PM Michael Goffioul <
> michael.goffi...@gmail.com> wrote:
>
>> I'm still experimenting with android-r-preview-2, and very close to the
>> end of the build process, I'm getting the following error:
>>
>> FAILED:
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.vdex
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.vdex
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.vdex
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.vdex
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.vdex
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.art
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.invocation
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.oat
>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.vdex
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-apache-xml.oat
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-bouncycastle.oat
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-core-icu4j.oat
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-core-libart.oat
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-okhttp.oat
>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot.oat
>> dex2oatd F 03-30 18:38:12  5762  5762 class_linker.cc:8451] Check failed:
>> previous_size == size (previous_size=108, size=104) Ljava/lang/DexCache;
>> Runtime aborting...
>>
>> Any hint would be appreciated. I've never had that error in previous
>> Android versions, I'm not sure what to look for.
>>
>> Note that the above error is accompanied with a message to add
>> "--runtime-arg -verbose:verifier" to the
>>
>
> My apologies, I hit send too early. I just meant to say the additional
> flag didn't provide more information.
>
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LvPSTaO7YEXLvorHOcKd%3DEi778-XPJyKC4E2YW2S7Hscw%40mail.gmail.com.

Re: [android-building] Re: Dex2oat failed to compile a boot image

2020-03-31 Thread Michael Goffioul 
Indeed, although there's a tagged manifest for android-r-preview-2 (
https://android.googlesource.com/platform/manifest/+/d28565b32bcdbdff11e3f47649032fb0a02d7828
<https://android.googlesource.com/platform/manifest/+/d28565b32bcdbdff11e3f47649032fb0a02d7828%5E%21/#F0>)
and all all repos have the corresponding tag, this is completely useless,
as some core repos are out of sync or incorrectly tagged. In particular,
all my last build errors were due to inconsistent "art" and "libcore"
repositories: various changes covering both repos had their "art" part
included in the tag android-r-preview-2, while their "libcore" part was
left ouf of the same tag.

On Tue, Mar 31, 2020 at 1:45 AM 'Dan Willemsen' via Android Building <
android-building@googlegroups.com> wrote:

> As posted with the announcement:
>
> This preview is not a full platform update and only for reference.
>
>
> I would not expect this to build or necessarily get any further (eg
> boot/etc). Only some components have been updated, and that's done on a
> git-project boundary, so the entire build graph gets skewed significantly
> between old and new projects.
>
> - Dan
>
> On Mon, Mar 30, 2020 at 10:10 PM Guoyin Chen 
> wrote:
>
>> There are quite a lot of build error with android-preview-2 build. I
>> think google have several wrong tag on different gits.
>>
>> On Tuesday, March 31, 2020 at 8:41:18 AM UTC+8, Michael Goffioul wrote:
>>>
>>> I found out I was missing this change:
>>> https://android.googlesource.com/platform/libcore/+/271b4054b825efca3176c4ac11565e33cff5d09a
>>> I cherry-picked it, but now I'm getting a different error:
>>>
>>> dex2oatd F 03-30 19:11:33  6962  6962 unstarted_runtime.cc:2085] Calling
>>> native method void android.system.OsConstants.initConstants() in an
>>> unstarted non-transactional runtime
>>>
>>>
>>> On Mon, Mar 30, 2020 at 6:59 PM Michael Goffioul 
>>> wrote:
>>>
>>>> On Mon, Mar 30, 2020 at 6:52 PM Michael Goffioul 
>>>> wrote:
>>>>
>>>>> I'm still experimenting with android-r-preview-2, and very close to
>>>>> the end of the build process, I'm getting the following error:
>>>>>
>>>>> FAILED:
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-apache-xml.vdex
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-bouncycastle.vdex
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-icu4j.vdex
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-core-libart.vdex
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot-okhttp.vdex
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.art
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.invocation
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.oat
>>>>> out/soong/x86/dex_artjars/android/apex/com.android.art/javalib/x86/boot.vdex
>>>>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-apache-xml.oat
>>>>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-bouncycastle.oat
>>>>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-core-icu4j.oat
>>>>> out/soong/x86/dex_artjars_unstripped/android/apex/com.android.art/javalib/x86/boot-core-libart.oat
>

[android-building] Cannot build android-platform-12.1.0_r4

2022-06-14 Thread Michael Goffioul 
Soong fails with the following error:

error: frameworks/opt/car/setupwizard/library/main/Android.bp:20:1:
"car-setup-wizard-lib" depends on undefined module
"androidx.car_car-resources-partially-dejetified"

>From what I can see, the code in packages/apps/Car/libs/androidx-car/ has
been moved to packages/apps/Car/systemlibs/ [1], but that repo is not
included in the manifest corresponding to the same tag [2], nor does that
repo have a corresponding tag [3].

[1]
https://android.googlesource.com/platform/packages/apps/Car/libs/+/3c40ccd5c8ae39a10c3a712a81f921b30ad687e5
[2]
https://android.googlesource.com/platform/manifest/+/375232ef8feb2bea64c31b39efc000db977b7af4/default.xml
[3] https://android.googlesource.com/platform/packages/apps/Car/systemlibs/

Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Lun%2BwxtkE9NXY9yhwUg2nXZLYHUuGMaiRrmCFE7PB9v9w%40mail.gmail.com.

Re: [android-building] Cannot build android-platform-12.1.0_r4

2022-06-15 Thread Michael Goffioul 
Hi Al,

I think you're confusing tags android-12.1.0_r4 and
android-platform-12.1.0_r4. I'm talking about the latter. For the
"-platform" tag, the code has indeed been moved to its new location, but
the corresponding repo is not included in the manifest (nor does it have a
corresponding tag anyway).

Michael.


On Wed, Jun 15, 2022 at 5:56 PM Al Sutton  wrote:

> Hi Michael,
>
>
>
> The systemlibs seems to be a move for 12.1.0_r7+, which is why I wouldn’t
> expect it to be referenced in the r4. In r4 that library comes from a
> different repo[1], which is referenced in the r4 manifest[2].
>
>
>
> [1]
> https://android.googlesource.com/platform/packages/apps/Car/libs/+/refs/tags/android-12.1.0_r4/androidx-car/Android.bp
>
> [2]
> https://android.googlesource.com/platform/manifest/+/refs/heads/android-12.1.0_r4/default.xml#832
>
>
>
> Al.
>
>
>
>
>
> *From: *Michael Goffioul 
> *Sent: *14 June 2022 19:09
> *To: *android-building@googlegroups.com
> *Subject: *[android-building] Cannot build android-platform-12.1.0_r4
>
>
>
> Soong fails with the following error:
>
>
>
> error: frameworks/opt/car/setupwizard/library/main/Android.bp:20:1:
> "car-setup-wizard-lib" depends on undefined module
> "androidx.car_car-resources-partially-dejetified"
>
>
>
> From what I can see, the code in packages/apps/Car/libs/androidx-car/ has
> been moved to packages/apps/Car/systemlibs/ [1], but that repo is not
> included in the manifest corresponding to the same tag [2], nor does that
> repo have a corresponding tag [3].
>
>
>
> [1]
> https://android.googlesource.com/platform/packages/apps/Car/libs/+/3c40ccd5c8ae39a10c3a712a81f921b30ad687e5
>
> [2]
> https://android.googlesource.com/platform/manifest/+/375232ef8feb2bea64c31b39efc000db977b7af4/default.xml
>
> [3]
> https://android.googlesource.com/platform/packages/apps/Car/systemlibs/
>
>
>
> Michael.
>
>
>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/CAB-99Lun%2BwxtkE9NXY9yhwUg2nXZLYHUuGMaiRrmCFE7PB9v9w%40mail.gmail.com
> <https://groups.google.com/d/msgid/android-building/CAB-99Lun%2BwxtkE9NXY9yhwUg2nXZLYHUuGMaiRrmCFE7PB9v9w%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>
>
>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/1MzR0i-1noD4C1gDA-00vKp8%40mrelayeu.kundenserver.de
> <https://groups.google.com/d/msgid/android-building/1MzR0i-1noD4C1gDA-00vKp8%40mrelayeu.kundenserver.de?utm_medium=email&utm_source=footer>
> .
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Ls2je2HY%3DtRb9Uj-95t8DL48OnC1YVJrYAX_urLpRw5Ew%40mail.gmail.com.

[android-building] Re: dex2oatd fails due to unexpected mmap address

2022-07-28 Thread Michael Goffioul 
I forgot to mention the Android version is 7.1.2.

On Tue, Jul 26, 2022 at 2:34 PM Michael Goffioul 
wrote:

> Hi,
>
> I've been compiling an arm64 firmware (for AMLogic chipset) on the same
> Fedora-based system for a couple of years now. Today, after rebuilding from
> scratch, I faced the following error. Would anybody have any insight on
> what is going on?
>
> [100% 2/2] build
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> FAILED: /bin/bash -c "(mkdir -p
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ &&
> rm -f
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> ) && (find out/target/common/obj/APPS/AppInstaller_intermediates/ -maxdepth
> 1 -name \"classes*.dex\" | sort | xargs zip -qjX
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> ) && (mv
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
> ) && (rm -f
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> ) && (mkdir -p
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ ) &&
> (ANDROID_LOG_TAGS=\"*:e\" out/host/linux-x86/bin/dex2oatd --runtime-arg
> -Xms64m --runtime-arg -Xmx512m --runtime-arg -classpath --runtime-arg \"&\"
> --boot-image=out/target/product/p212/dex_bootjars/system/framework/boot.art
> --dex-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
> --dex-location=/system/app/AppInstaller/AppInstaller.apk
> --oat-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
> --android-root=out/target/product/p212/system --instruction-set=arm64
> --instruction-set-variant=generic --instruction-set-features=default
> --include-patch-information --runtime-arg -Xnorelocate
> --no-generate-debug-info --abort-on-hard-verifier-error
> --no-inline-from=core-oj.jar   --compile-pic ) && (rm
> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
> )"
> dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Could not create
> image space with image file
> 'out/target/product/p212/dex_bootjars/system/framework/boot.art'.
> Attempting to fall back to imageless running. Error was: Failed to load
> /system image
> 'out/target/product/p212/dex_bootjars/system/framework/arm64/boot-framework.art':
> Failed to mmap at expected address, mapped at 0x7fdaa8a0 instead of
> 0x70514000
> dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Attempted image:
> out/target/product/p212/dex_bootjars/system/framework/boot-framework.art
> dex2oatd E 449109 449109 art/runtime/runtime.cc:1050] Dex file fallback
> disabled, cannot continue without image.
> dex2oatd E 449109 449109 art/dex2oat/dex2oat.cc:2232] Failed to create
> runtime
> ninja: build stopped: subcommand failed.
>
> Thanks,
> Michael.
>
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LuOaHrge2BMLAwKiiKHRn-dhavbtQ8z2rYxbZQqw%3DnQUg%40mail.gmail.com.

[android-building] dex2oatd fails due to unexpected mmap address

2022-07-28 Thread Michael Goffioul 
Hi,

I've been compiling an arm64 firmware (for AMLogic chipset) on the same
Fedora-based system for a couple of years now. Today, after rebuilding from
scratch, I faced the following error. Would anybody have any insight on
what is going on?

[100% 2/2] build
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
FAILED: /bin/bash -c "(mkdir -p
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ &&
rm -f
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
) && (find out/target/common/obj/APPS/AppInstaller_intermediates/ -maxdepth
1 -name \"classes*.dex\" | sort | xargs zip -qjX
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
) && (mv
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
) && (rm -f
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
) && (mkdir -p
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ ) &&
(ANDROID_LOG_TAGS=\"*:e\" out/host/linux-x86/bin/dex2oatd --runtime-arg
-Xms64m --runtime-arg -Xmx512m --runtime-arg -classpath --runtime-arg \"&\"
--boot-image=out/target/product/p212/dex_bootjars/system/framework/boot.art
--dex-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
--dex-location=/system/app/AppInstaller/AppInstaller.apk
--oat-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
--android-root=out/target/product/p212/system --instruction-set=arm64
--instruction-set-variant=generic --instruction-set-features=default
--include-patch-information --runtime-arg -Xnorelocate
--no-generate-debug-info --abort-on-hard-verifier-error
--no-inline-from=core-oj.jar   --compile-pic ) && (rm
out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
)"
dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Could not create image
space with image file
'out/target/product/p212/dex_bootjars/system/framework/boot.art'.
Attempting to fall back to imageless running. Error was: Failed to load
/system image
'out/target/product/p212/dex_bootjars/system/framework/arm64/boot-framework.art':
Failed to mmap at expected address, mapped at 0x7fdaa8a0 instead of
0x70514000
dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Attempted image:
out/target/product/p212/dex_bootjars/system/framework/boot-framework.art
dex2oatd E 449109 449109 art/runtime/runtime.cc:1050] Dex file fallback
disabled, cannot continue without image.
dex2oatd E 449109 449109 art/dex2oat/dex2oat.cc:2232] Failed to create
runtime
ninja: build stopped: subcommand failed.

Thanks,
Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LvD%3DC%3DP4En7RzAq1JZ8Sb5jJND%2Bi9PsWf%2B51adD%3DfEWWg%40mail.gmail.com.

[android-building] Re: dex2oatd fails due to unexpected mmap address

2022-07-28 Thread Michael Goffioul 
This seems to be due to the kernel update on the build host, which switched
from 5.17 to 5.18. Because the error is related to unexpected mmap address,
and there was a recent kernel update on the Fedora build host, and after
suspecting everything else, I ended up suspecting the kernel itself. I
rebooted into the previous 5.17 kernel, and the error went away.

It would be nice if the Android devs could come up with a workaround that
does not involve not upgrading the build host kernel. I'm happy to test any
proposed solution.

Thanks,
Michael.


On Tue, Jul 26, 2022 at 2:36 PM Michael Goffioul 
wrote:

> I forgot to mention the Android version is 7.1.2.
>
> On Tue, Jul 26, 2022 at 2:34 PM Michael Goffioul <
> michael.goffi...@gmail.com> wrote:
>
>> Hi,
>>
>> I've been compiling an arm64 firmware (for AMLogic chipset) on the same
>> Fedora-based system for a couple of years now. Today, after rebuilding from
>> scratch, I faced the following error. Would anybody have any insight on
>> what is going on?
>>
>> [100% 2/2] build
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> FAILED: /bin/bash -c "(mkdir -p
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ &&
>> rm -f
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> ) && (find out/target/common/obj/APPS/AppInstaller_intermediates/ -maxdepth
>> 1 -name \"classes*.dex\" | sort | xargs zip -qjX
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> ) && (mv
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
>> ) && (rm -f
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> ) && (mkdir -p
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/ ) &&
>> (ANDROID_LOG_TAGS=\"*:e\" out/host/linux-x86/bin/dex2oatd --runtime-arg
>> -Xms64m --runtime-arg -Xmx512m --runtime-arg -classpath --runtime-arg \"&\"
>> --boot-image=out/target/product/p212/dex_bootjars/system/framework/boot.art
>> --dex-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
>> --dex-location=/system/app/AppInstaller/AppInstaller.apk
>> --oat-file=out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex
>> --android-root=out/target/product/p212/system --instruction-set=arm64
>> --instruction-set-variant=generic --instruction-set-features=default
>> --include-patch-information --runtime-arg -Xnorelocate
>> --no-generate-debug-info --abort-on-hard-verifier-error
>> --no-inline-from=core-oj.jar   --compile-pic ) && (rm
>> out/target/product/p212/obj/APPS/AppInstaller_intermediates/oat/arm64/package.odex.input
>> )"
>> dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Could not create
>> image space with image file
>> 'out/target/product/p212/dex_bootjars/system/framework/boot.art'.
>> Attempting to fall back to imageless running. Error was: Failed to load
>> /system image
>> 'out/target/product/p212/dex_bootjars/system/framework/arm64/boot-framework.art':
>> Failed to mmap at expected address, mapped at 0x7fdaa8a0 instead of
>> 0x70514000
>> dex2oatd E 449109 449109 art/runtime/gc/heap.cc:324] Attempted image:
>> out/target/product/p212/dex_bootjars/system/framework/boot-framework.art
>> dex2oatd E 449109 449109 art/runtime/runtime.cc:1050] Dex file fallback
>> disabled, cannot continue without image.
>> dex2oatd E 449109 449109 art/dex2oat/dex2oat.cc:2232] Failed to create
>> runtime
>> ninja: build stopped: subcommand failed.
>>
>> Thanks,
>> Michael.
>>
>>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LsRWE%3D62Q9VOVD9ok4oSvwVdUYCB%3D7NRnXn_QM-QjMh2g%40mail.gmail.com.

Re: [android-building] Re: dex2oatd fails due to unexpected mmap address

2022-08-24 Thread Michael Goffioul 
On Thu, Jul 28, 2022 at 8:58 PM 'Dan Willemsen' via Android Building <
android-building@googlegroups.com> wrote:

> Have you tried to reproduce with a newer version of Android? Android 7 is
> about 6 years old, so I wouldn't be expecting any changes to it.
>
> You may want to talk to your distribution, this could be a configuration
> change during the kernel upgrade. Breaking existing binaries is something
> the kernel generally attempts to avoid.
>

I agree. So the most probable conclusion is that the problem comes from the
Android side, not the kernel/distribution side.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LurxUpamFW9eCZf4fwvZg72Pk9%3DnUB%3DYd4Gs1reX%3DDnvw%40mail.gmail.com.

Re: [android-building] Re: dex2oatd fails due to unexpected mmap address

2022-08-24 Thread Michael Goffioul 
On Sun, Aug 14, 2022 at 8:30 AM Michael Goffioul 
wrote:

>
>
> On Thu, Jul 28, 2022 at 8:58 PM 'Dan Willemsen' via Android Building <
> android-building@googlegroups.com> wrote:
>
>> Have you tried to reproduce with a newer version of Android? Android 7 is
>> about 6 years old, so I wouldn't be expecting any changes to it.
>>
>> You may want to talk to your distribution, this could be a configuration
>> change during the kernel upgrade. Breaking existing binaries is something
>> the kernel generally attempts to avoid.
>>
>
> I agree. So the most probable conclusion is that the problem comes from
> the Android side, not the kernel/distribution side.
>
>
After adding debug lines in dex2oatd source code, it appears the error
comes from this line:
https://cs.android.com/android/platform/superproject/+/android-7.1.2_r1:art/runtime/mem_map.cc;l=891
where addr=0x70514000 and flags=0002. Correct me if I'm wrong, but
given that the mmap flags do not include MAP_FIXED, the provided address is
only interpreted as a hint, there's no guarantee that the mapped address
will be the one provided as argument.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LvRNThpX3m%2BRwAMfZiUwOnFtUSMbsOY%2B80eDc5pO%3DtowQ%40mail.gmail.com.

[android-building] Cannot build Android 13 with audio.a2dp.default module

2022-09-13 Thread Michael Goffioul 
I'm using a source tree derived from manifest android-13.0.0_r4. The build
also adds audio.a2dp.default to PRODUCT_PACKAGES. The ckati step then fails
with the following error:

build/make/core/tasks/platform_availability_check.mk:37: warning:
 Following modules are requested to be installed. But are not available for
platform because they do
 not have "//apex_available:platform" or they depend on other modules that
are not available for platform
Offending entries:
audio.a2dp.default:packages/modules/Bluetooth/system/audio_a2dp_hw

Could you tell me how to avoid that error?

Note that it seems I'm not the only one having that problem, I found this
gerrit change that refers the exact same problem:
https://android-review.googlesource.com/c/device/ti/beagle-x15/+/1919398

Thanks,
Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LsR412BWPwztyQ5e-%3DV%3DYBHbSA6-0pOYqCAn0pTeW64gg%40mail.gmail.com.

Re: [android-building] Cannot build Android 13 with audio.a2dp.default module

2022-09-14 Thread Michael Goffioul 
Hi Dan,

Thanks for the anwer. Would mind elaborate what you mean with "embedded in
the mainline module"? When removing module audio.a2dp.default from
PRODUCT_PACKAGES, the build succeeds, but I don't see the module (
audio.a2dp.default.so) anywhere on the image. I also don't see
"audio.a2dp.default" referenced in any Android.bp on https://cs.android.com.
But the audio HAL 7.1 is still looking for that module.

Have the functionalities of audio.a2dp.default been integrated elsewhere,
making the module obsolete?

Thanks,
Michael.


On Wed, Sep 14, 2022 at 12:27 AM 'Dan Willemsen' via Android Building <
android-building@googlegroups.com> wrote:

> audio.a2dp.default is now part of the bluetooth mainline module
> <https://source.android.com/docs/core/architecture/modular-system>, and
> cannot be installed independently. It will be embedded in the mainline
> module instead.
>
> - Dan
>
> On Tue, Sep 13, 2022 at 6:11 PM Michael Goffioul <
> michael.goffi...@gmail.com> wrote:
>
>> I'm using a source tree derived from manifest android-13.0.0_r4. The
>> build also adds audio.a2dp.default to PRODUCT_PACKAGES. The ckati step then
>> fails with the following error:
>>
>> build/make/core/tasks/platform_availability_check.mk:37: warning:
>>  Following modules are requested to be installed. But are not available for
>> platform because they do
>>  not have "//apex_available:platform" or they depend on other modules
>> that are not available for platform
>> Offending entries:
>> audio.a2dp.default:packages/modules/Bluetooth/system/audio_a2dp_hw
>>
>> Could you tell me how to avoid that error?
>>
>> Note that it seems I'm not the only one having that problem, I found this
>> gerrit change that refers the exact same problem:
>> https://android-review.googlesource.com/c/device/ti/beagle-x15/+/1919398
>>
>> Thanks,
>> Michael.
>>
>> --
>> --
>> You received this message because you are subscribed to the "Android
>> Building" mailing list.
>> To post to this group, send email to android-building@googlegroups.com
>> To unsubscribe from this group, send email to
>> android-building+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/android-building?hl=en
>>
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "Android Building" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to android-building+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/android-building/CAB-99LsR412BWPwztyQ5e-%3DV%3DYBHbSA6-0pOYqCAn0pTeW64gg%40mail.gmail.com
>> <https://groups.google.com/d/msgid/android-building/CAB-99LsR412BWPwztyQ5e-%3DV%3DYBHbSA6-0pOYqCAn0pTeW64gg%40mail.gmail.com?utm_medium=email&utm_source=footer>
>> .
>>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/CALQgHd%3DeCFRWz4xqV8iR9GGTkMaqgwGABsKkuo_f78JXZ-RPLw%40mail.gmail.com
> <https://groups.google.com/d/msgid/android-building/CALQgHd%3DeCFRWz4xqV8iR9GGTkMaqgwGABsKkuo_f78JXZ-RPLw%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Lt5dBPSrpxzOAK1_NcyfCHJq_LEDzLhMoWEE7HBC5o_tg%40mail.gmail.com.

Re: [android-building] Re: Android 13 Released - main.mk:2011:error:missing separator

2022-09-14 Thread Michael Goffioul 
It turned out the problem was caused by the presence of a file
in PRODUCT_COPY_FILES that contains a comma in its filename. This
interferes with the macro expansion of make. We'll have to test the
proposed commit, but from a first inspection, I think the problem will
still be present, as the offending macro [1] has just been relocated
elsewhere

[1]
https://cs.android.com/android/platform/superproject/+/android-13.0.0_r3:build/make/core/definitions.mk;l=665


On Wed, Sep 14, 2022 at 2:35 PM Peter Yoon  wrote:

> Try  applying following change.
>
>
> https://android.googlesource.com/platform/build/+/ba3372439f34a570af02511a438c6e3c947ce3f1
>
> On Monday, August 29, 2022 at 10:51:33 AM UTC+9 issor...@gmail.com wrote:
>
>>
>> Just a question, do you also get the following building error when using
>> platform/build project at tag android-13.0.0_r3 ?
>>
>> [ 99% 210/211] finishing build rules ...
>> FAILED: build/make/core/main.mk:2011: error: missing separator.
>> 08:59:48 ckati failed with: exit status 1
>>  failed to build some targets (03:38 (mm:ss)) 
>>
>> It seams related to some violation of gnu make syntax, is there some way
>> to avoit it?
>> Thanks
>>
> --
> --
> You received this message because you are subscribed to the "Android
> Building" mailing list.
> To post to this group, send email to android-building@googlegroups.com
> To unsubscribe from this group, send email to
> android-building+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/android-building?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Android Building" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-building+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/android-building/f68004c9-f37f-41ac-b887-5a20192c238cn%40googlegroups.com
> 
> .
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Lv8eb_5usWpY%3DXyUBCC4McUX5Rp%2Bn9uF%2B3jThROCZ1DiQ%40mail.gmail.com.

[android-building] Re: External camera service for 64 bits

2023-01-10 Thread Michael Goffioul 
Well, it turns out I believe there's invalid code in the external camera
provider HAL, which happens to work only on 32 bits. More specifically,
this snippet:
https://cs.android.com/android/platform/superproject/+/master:hardware/interfaces/camera/provider/2.4/default/ExternalCameraProviderImpl_2_4.cpp;l=44-47

const char* kDevicePath = "/dev/";
constexpr char kPrefix[] = "video";
constexpr int kPrefixLen = sizeof(kPrefix) - 1;
constexpr int kDevicePrefixLen = sizeof(kDevicePath) + kPrefixLen + 1;

The above constants are used in the code to locate the numerical value in
the video device name, e.g. "2" in "/dev/video2". The constant kPrefixLen
is computed from constant char array kPrefix, minus one, as the array will
have the terminating null byte (which doesn't count when computing the
digit offset), that is 5. But then, kDevicePrefixLen does not make sense:
- first it uses "sizeof(kDevicePath)", which is the size of a pointer; that
is 4 on 32 bits, 8 on 64 bits
- second it adds one to the invalidly computed length of "/dev/video" (I'm
assuming that was the intent of the author)
Assuming the author thought "sizeof(kDevicePath)" would yield the length of
the constant string "/dev/" (I think that a reasonable assumption), then
incrementing by one "sizeof(kDevicePath) + kPrefixLen" does not make sense,
as the non-incremented value is what you're after. But on 32 bits,
something funny happens: the length of "/dev/" is 5, and
sizeof(kDevicePath) is 4. Maybe the author noticed that kDevicePrefixLen
didn't have the expected value, then noticed the previous line where
kPrefixLen is decremented by one, thought kPrefixLen should then be
re-incremented, and Bob's your uncle. Works fine on 32 bits, falls flat on
64 bits.

Not sure if anybody at Google is reading this, but you might wanna have a
look at that more closely.

Michael.




On Thu, Sep 29, 2022 at 3:09 PM Michael Goffioul 
wrote:

> Is there a particular reason to keep the external camera service a 32-bit
> only component [1]? From what I can see, the cameraserver process has been
> made available for 64-bit build [2].
>
> Thanks,
> Michael.
>
> [1]
> https://cs.android.com/android/platform/superproject/+/master:hardware/interfaces/camera/provider/2.7/default/Android.bp;l=62
> [2]
> https://cs.android.com/android/_/android/platform/frameworks/av/+/642fc1d0ef969c29a8882895d992b008b3bb7dd1
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Lt_g9A_vA0jrY4fb-ACJH-sWWQZJwEny_AmCSwPjtHmjQ%40mail.gmail.com.

Re: [android-building] Re: External camera service for 64 bits

2023-01-12 Thread Michael Goffioul 
On Wed, Jan 11, 2023 at 2:32 AM 'enh' via Android Building <
android-building@googlegroups.com> wrote:

> On Tue, Jan 10, 2023 at 5:17 PM Michael Goffioul <
> michael.goffi...@gmail.com> wrote:
>
>> Well, it turns out I believe there's invalid code in the external camera
>> provider HAL, which happens to work only on 32 bits. More specifically,
>> this snippet:
>> https://cs.android.com/android/platform/superproject/+/master:hardware/interfaces/camera/provider/2.4/default/ExternalCameraProviderImpl_2_4.cpp;l=44-47
>>
>> const char* kDevicePath = "/dev/";
>> constexpr char kPrefix[] = "video";
>> constexpr int kPrefixLen = sizeof(kPrefix) - 1;
>> constexpr int kDevicePrefixLen = sizeof(kDevicePath) + kPrefixLen + 1;
>>
>> The above constants are used in the code to locate the numerical value in
>> the video device name, e.g. "2" in "/dev/video2". The constant kPrefixLen
>> is computed from constant char array kPrefix, minus one, as the array will
>> have the terminating null byte (which doesn't count when computing the
>> digit offset), that is 5. But then, kDevicePrefixLen does not make sense:
>> - first it uses "sizeof(kDevicePath)", which is the size of a pointer;
>> that is 4 on 32 bits, 8 on 64 bits
>>
>
> no, that's weird -- and i tell people not to do it in code review because
> it's confusing -- but it's correct because it's an _array_ rather than a
> char*.
>

With all due respect, I think you may be mistaken here. The constant
kDevicePath is not declared as an array, but as a const char*, so
sizeof(kDevicePath) is the size of a pointer. E.g. try to compile the
following in 32 or 64 bit:

#include 
const char* kDevicePath_1 = "/dev/";
constexpr char kDevicePath_2[] = "/dev/";
int main(int argc, char** argv) {
std::cout << "sizeof(kDevicePath_1) = " << sizeof(kDevicePath_1) <<
std::endl;
std::cout << "sizeof(kDevicePath_2) = " << sizeof(kDevicePath_2) <<
std::endl;
return 0;
}

I get the following results (note: I've tried with both gcc - from Fedora
37 - and clang - from AOSP prebuilts @ android-13.0.0_r24):

$ g++ -m32 -o testsize testsize.cpp
$ ./testsize
sizeof(kDevicePath_1) = 4
sizeof(kDevicePath_2) = 6
$ g++ -o testsize testsize.cpp
$ ./testsize
sizeof(kDevicePath_1) = 8
sizeof(kDevicePath_2) = 6



> (it's actually 6 in both cases, because it also includes the '\0', which
> is one reason why people use it rather than the more obvious `strlen() +
> 1`.)
>

But let's still assume I'm wrong, and that `sizeof(kDevicePath)` does
indeed yield 6, like you said. Then kDevicePrefixLen would be 12. However
the wanted value is the length of the string "/dev/video", that is 10. That
is, kDevicePrefixLen would have an incorrect value.


>  what's the actual failure you're seeing?
>

First let me stress the fact that I'm compiling the external camera
provider HAL in 64 bits mode (I'm targeting a 64-bits only build). I have 2
USB cameras attached to the device, using /dev/video0 and /dev/video2. Both
devices are reported to cameraserver with the same identifier
device@3.4/external/0,
so basically cameraserver only sees 1 camera device. I believe that's due
this snippet in ExternalCameraProviderImpl_2_4::addExternalCamera:

std::string cameraId = std::to_string(mCfg.cameraIdOffset +
  std::atoi(devName +
kDevicePrefixLen));

and the fact that std::atoi returns 0 if no conversion can be made.

As a side note, the above snippet also performs an out-of-bound array
access on 64 bits platform. Granted, the external camera service is
currently forced into 32 bits compilation in AOSP, so this out-of-bound
access is only latent.

Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LvLjaNO3yFP%3Dt7MuP6BAo71UqyEictSpCdVWWOXNZvhyA%40mail.gmail.com.

[android-building] Is AVX2 now a mandatory ABI requirement for x86 platforms?

2023-11-14 Thread Michael Goffioul 
Since this commit, webrtc audio processing module is compiled with AVX2
enabled.

https://cs.android.com/android/_/android/platform/external/webrtc/+/b5b78192e10ba90978fb8228e7eb9fcf75c83f0c

This makes the module fail to load on platforms that do not support AVX2.
E.g. generates SIGILL at a location like this one:

https://cs.android.com/android/platform/superproject/main/+/main:external/webrtc/modules/audio_processing/include/audio_processing.h;drc=4366c5469fefb6a1e641378a2bb152e3c5ad7421;l=186

Is this ABI requirement change documented anywhere? The main info about ABI
I can find is this one, and it doesn't say AVX2 is mandatory (you can use
it, but you need the proper safeguard, which webrtc module hasn't):

https://developer.android.com/ndk/guides/abis

Thanks,
Michael.

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99LuRHcyK%2BGrH8GSYwFWVCPhDNX7Xb7ZnoN62oFAzyagd3g%40mail.gmail.com.

Re: [android-building] Is AVX2 now a mandatory ABI requirement for x86 platforms?

2024-04-29 Thread Michael Goffioul 
Hello? Anybody out there?


On Tue, Dec 19, 2023 at 9:29 AM Michael Goffioul 
wrote:

> Any update on this? I guess that now that the cat's out of the bag, it'll
> be difficult to put it back in, and AVX2 is a de facto ABI requirement.
>
>
> On Tue, Nov 14, 2023 at 1:13 PM 'enh' via Android Building <
> android-building@googlegroups.com> wrote:
>
>> +the webrtc OWNERS. (because, no, the Android ABI does not include
>> AVX2. i'm curious why this hasn't caused trouble for folks using the
>> emulator on older hardware, since the Studio requirements also don't
>> include AVX2?)
>>
>> On Tue, Nov 14, 2023 at 9:46 AM Michael Goffioul
>>  wrote:
>> >
>> > Since this commit, webrtc audio processing module is compiled with AVX2
>> enabled.
>> >
>> >
>> https://cs.android.com/android/_/android/platform/external/webrtc/+/b5b78192e10ba90978fb8228e7eb9fcf75c83f0c
>> >
>> > This makes the module fail to load on platforms that do not support
>> AVX2. E.g. generates SIGILL at a location like this one:
>> >
>> >
>> https://cs.android.com/android/platform/superproject/main/+/main:external/webrtc/modules/audio_processing/include/audio_processing.h;drc=4366c5469fefb6a1e641378a2bb152e3c5ad7421;l=186
>> >
>> > Is this ABI requirement change documented anywhere? The main info about
>> ABI I can find is this one, and it doesn't say AVX2 is mandatory (you can
>> use it, but you need the proper safeguard, which webrtc module hasn't):
>> >
>> > https://developer.android.com/ndk/guides/abis
>> >
>> > Thanks,
>> > Michael.
>> >
>> > --
>> > --
>> > You received this message because you are subscribed to the "Android
>> Building" mailing list.
>> > To post to this group, send email to android-building@googlegroups.com
>> > To unsubscribe from this group, send email to
>> > android-building+unsubscr...@googlegroups.com
>> > For more options, visit this group at
>> > http://groups.google.com/group/android-building?hl=en
>> >
>> > ---
>> > You received this message because you are subscribed to the Google
>> Groups "Android Building" group.
>> > To unsubscribe from this group and stop receiving emails from it, send
>> an email to android-building+unsubscr...@googlegroups.com.
>> > To view this discussion on the web visit
>> https://groups.google.com/d/msgid/android-building/CAB-99LuRHcyK%2BGrH8GSYwFWVCPhDNX7Xb7ZnoN62oFAzyagd3g%40mail.gmail.com
>> .
>>
>> --
>> --
>> You received this message because you are subscribed to the "Android
>> Building" mailing list.
>> To post to this group, send email to android-building@googlegroups.com
>> To unsubscribe from this group, send email to
>> android-building+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/android-building?hl=en
>>
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "Android Building" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to android-building+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/android-building/CAJgzZoo%2BDn7SqLVpubhV_%3DPDtOdiu6aVa3W1Y_fT3p1Uhkwj0A%40mail.gmail.com
>> .
>>
>

-- 
-- 
You received this message because you are subscribed to the "Android Building" 
mailing list.
To post to this group, send email to android-building@googlegroups.com
To unsubscribe from this group, send email to
android-building+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-building?hl=en

--- 
You received this message because you are subscribed to the Google Groups 
"Android Building" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to android-building+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/android-building/CAB-99Lub7taSoa9K1FujaqLzJAEobcw8REWyrmEcpU27ufWKzg%40mail.gmail.com.