After bug 1002891 and bug 1003902 are fixed, I think we can label storage/src as FAIL_ON_WARNINGS.

Here's the patch. Verified locally (on Linux) that this breaks my build, if I don't have the "depends-on" bugs' patches (due to warning being treated as an error). Build succeeds if I apply those patches as well. So, this is effective.

Try run with this patch & "depends-on" bugs' patches: https://tbpl.mozilla.org/?tree=Try&rev=912bda3beb6f

I tried to FAIL_ON_WARNINGS storage/src a couple days ago: https://tbpl.mozilla.org/?tree=Try&rev=bd6c050df6df and (just) filed bug 1003921 for B2G Desktop OS X Opt build warning (using gcc 4.4, I believe): storage/src/mozStorageAsyncStatementExecution.cpp:456:18: error: ignoring return value of function declared with const attribute [-Werror,-Wunused-value] NS_WARN_IF(NS_FAILED(mConnection->rollbackTransactionInternal(mNativeConnection))); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ../../dist/include/nsError.h:183:40: note: expanded from macro 'NS_FAILED' #define NS_FAILED(_nsresult) ((bool)MOZ_UNLIKELY(NS_FAILED_impl(_nsresult))) ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ../../dist/include/mozilla/Likely.h:17:28: note: expanded from macro 'MOZ_UNLIKELY' # define MOZ_UNLIKELY(x) (__builtin_expect(!!(x), 0)) ^~~~~~~~~~~~~~~~ ~~~~~~~~ ../../dist/include/nsDebug.h:47:38: note: expanded from macro 'NS_WARN_IF' #define NS_WARN_IF(condition) (bool)(condition) ^

Comment on attachment 8415375 [details] [diff] [review] fix v1 Review of attachment 8415375 [details] [diff] [review]: ----------------------------------------------------------------- r=me provided we get a full try-build with also tests.

(In reply to Marco Bonardo [:mak] from comment #4) > r=me provided we get a full try-build with also tests. Thanks! Try run: https://tbpl.mozilla.org/?tree=Try&rev=02c0798aa406 Green, save for some unrelated-looking intermittent/apparently-intermittent (green after retrigger) oranges.

Landed: https://hg.mozilla.org/integration/mozilla-inbound/rev/628dd71d10ca