711034 - Native Fennec is still creating places.sqlite

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect)

Description

[Gian-Carlo Pascutto [:gcp]]

Native Fennec is supposed to use the Android DB or a local DB, but in fact the places.sqlite file is still created and the cache is eating up 0.5MB RAM.

Comment 1

[Marco Bonardo [:mak]]

Database::Init() should bail out, so I don't see how's that possible since should be the only code path initing the database.

Comment 2

[Gian-Carlo Pascutto [:gcp]]

As far as I can see, it's because configure.in doesn't contain what's needed to make MOZ_ANDROID_HISTORY actually take effect. Fixed that, seems to be OK now.

I also tried to build without MOZ_PLACES, but, well, I guess bug 702037 already explains that doesn't work.

Comment 3

[Gian-Carlo Pascutto [:gcp]]

Attachment: Patch 1. Make MOZ_ANDROID_HISTORY actually work

Comment 4

[Marco Bonardo [:mak]]

(In reply to Gian-Carlo Pascutto (:gcp) from comment #2)
> I also tried to build without MOZ_PLACES, but, well, I guess bug 702037
> already explains that doesn't work.

Can't work since Android history should be in its own component for that to work, as I said there.
Btw, thanks for fixing those missing bits.

Comment 5

[Doug Turner (:dougt)]

Comment on attachment 582007 [details] [diff] [review]
Patch 1. Make MOZ_ANDROID_HISTORY actually work

Review of attachment 582007 [details] [diff] [review]:
-----------------------------------------------------------------

missed this, i guess, when we merged birch over.

Comment 6

[Gian-Carlo Pascutto [:gcp]]

I'm now seeing:

I/Gecko   (23085): ###!!! ASSERTION: nsAndroidHistory not thread-safe: '_mOwningThread.GetThread() == PR_GetCurrentThread()', file /home/morbo/hg/mozilla-central/toolkit/components/places/nsAndroidHistory.cpp, line 44
I/Gecko   (23085): ###!!! ASSERTION: nsAndroidHistory not thread-safe: '_mOwningThread.GetThread() == PR_GetCurrentThread()', file /home/morbo/hg/mozilla-central/toolkit/components/places/nsAndroidHistory.cpp, line 44

I don't know the source - our Java code seems to be calling it properly on the Gecko thread.

Comment 7

[Gian-Carlo Pascutto [:gcp]]

https://tbpl.mozilla.org/?tree=Try&rev=18b7c05df167

https://hg.mozilla.org/integration/mozilla-inbound/rev/1bb4179d747e

Comment 8

[Matt Brubeck (:mbrubeck)]

https://hg.mozilla.org/mozilla-central/rev/1bb4179d747e