579186 - Mozilla-central N810 builds do not launch (GLIBCXX_3.4.9 not found)

Status: RESOLVED INCOMPLETE

(Firefox for Android Graveyard :: General, defect)

Description

[John Ford [:jhford] CET/CEST Berlin Time]

I downloaded a tinderbox-build of fennec for maemo4 built against trunk and am getting the message

/builds/fennec/fennec: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.9' not
found (required by /builds/fennec/libxul.so)

I am also getting 100% of tests failing on the device with the message:

INFO | automation.py | Application pid: 1602
/builds/fennec/fennec: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.9' not
found (required by /builds/fennec/libxul.so)
TEST-UNEXPECTED-FAIL | automation.py | Exited with code 1 during test run
INFO | automation.py | Application ran for: 0:00:00.195434
INFO | automation.py | Reading PID log: /var/tmp/tmpeC37hipidlog
WARNING | automationutils.processLeakLog() | refcount logging is off, so leaks
can't be detected!

REFTEST INFO | runreftest.py | Running tests: end.

Unfortunately, the results of the test runs are green, so it is very difficult to find the regression range.

I have filed bug 579184 to make sure that the green test results on failing runs doesn't happen again in the future.

N900 builds launch and run fine.

Comment 1

[Nick Thomas [:nthomas] (UTC+12)]

That looks very much like Vlad's ANGLE patch requiring a later version of libstdc++ than is present on the n810s.

Comment 2

[John Ford [:jhford] CET/CEST Berlin Time]

I guess this means that we either need to start shipping our own libstdc++, disable that patch for maemo4 builds or stop building maemo4 for mozilla-central.

Comment 3

[John Ford [:jhford] CET/CEST Berlin Time]

It looks like the SDK has a version of libstdc++ that does have this string (6.0.9) but the devices don't actually have that version.  The devices and the SDK both have 6.0.3, which does not have this string.

I wonder if we could build forcing linking to libstdc++.so.6.0.3.  If we were to do this, I wonder if this is going to work with the ANGLE patch.


[sbox-FREMANTLE_ARMEL: /usr/lib] > strings libstdc++.so.6.0.9  | grep GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_3.4.4
GLIBCXX_3.4.5
GLIBCXX_3.4.6
GLIBCXX_3.4.7
GLIBCXX_3.4.8
GLIBCXX_3.4.9
GLIBCXX_FORCE_NEW

[sbox-CHINOOK-ARMEL-2007: /usr/lib] > strings libstdc++.so.6.0.9 | grep GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_3.4.4
GLIBCXX_3.4.5
GLIBCXX_3.4.6
GLIBCXX_3.4.7
GLIBCXX_3.4.8
GLIBCXX_3.4.9
GLIBCXX_FORCE_NEW
[sbox-CHINOOK-ARMEL-2007: /usr/lib] > strings libstdc++.so.6.0.3 | grep GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_FORCE_NEW

maemo-n810-31:~# ls /usr/lib/libstdc\+\+.so.6*
/usr/lib/libstdc++.so.6      /usr/lib/libstdc++.so.6.0.3

Comment 4

[Mark Finkle (:mfinkle) (use needinfo?)]

N810/maemo4 is definitely a lower priority going forward, but being able to build and run should be a goal. What would statically linking a libstdc++ entail?

Comment 5

[Aki Sasaki (not active)]

Per bug 578880, this might be angle: https://bugzilla.mozilla.org/show_bug.cgi?id=578880#c5

Comment 6

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

Met Stuart, he'll look at it.

Comment 7

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to comment #4)
> N810/maemo4 is definitely a lower priority going forward, but being able to
> build and run should be a goal. What would statically linking a libstdc++
> entail?

I am not sure how to do that specifically.

I just copied libstdc++.so.6.0.9 to /usr/lib on the phone and changed the /usr/lib/libstdc++.so.6 to point to 6.0.9

This resulted in a fennec that launched.  I did some extremely limited testing and it seems like it is stable.

The source of the stray v6.0.9 library seems to be related to us using the cs2007q3-glibc2.5-arm7 toolchain instead of cs2005q3.2-glibc-arm.

Comment 8

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to comment #5)
> Per bug 578880, this might be angle:
> https://bugzilla.mozilla.org/show_bug.cgi?id=578880#c5

yes, but the solution that we have for that platform will not work on device without us distributing our own copy of libstdc++ aiui

Comment 9

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to comment #6)
> Met Stuart, he'll look at it.

Fixed missing assignment.

Comment 10

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to comment #9)
> (In reply to comment #6)
> > Met Stuart, he'll look at it.
> 
> Fixed missing assignment.

Ping - any update?(In reply to comment #6)

Comment 11

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

summary of quick meeting with vlad/stuart/joduinn:

stuart/vlad will get change the build config to remove dependency on this library. Possible ETA this week sometime.

Comment 12

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to comment #11)
> summary of quick meeting with vlad/stuart/joduinn:
> 
> stuart/vlad will get change the build config to remove dependency on this
> library. Possible ETA this week sometime.

stuart: ping?

Comment 13

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: disable on maemo5

Comment 14

[Aki Sasaki (not active)]

N810s were EOLed long before the N900s, which are now EOLed.