Created attachment 215979 [details] [review]
0001-ephy-profile-utils-fix-memory-leak.patch

Created attachment 215980 [details] [review]
0002-ephy-bookmark-action-fix-memory-leak.patch

Created attachment 215981 [details] [review]
0003-ephy-bookmarks-fix-memory-leaks.patch

Created attachment 215982 [details] [review]
0004-ephy-web-view-fix-GList-leak.patch

Created attachment 215983 [details] [review]
0005-ephy-session-fix-memory-leak-in-ephy_session_save.patch

Created attachment 215984 [details] [review]
0006-ephy-completion-model-fix-GList-of-EphyHistoryURL-le.patch

Created attachment 216099 [details] [review]
0007-ephy-bookmarks-editor-fix-memory-leak.patch

Created attachment 216100 [details] [review]
0008-ephy-bookmarks-editor-fix-GList-leaks.patch

Review of attachment 215979 [details] [review]:

Good catch.

Review of attachment 215980 [details] [review]:

LRTM

Review of attachment 215981 [details] [review]:

Maybe you can merge it with the previous one, but otherwise looks right.

Review of attachment 215982 [details] [review]:

Looks about right, too.

Review of attachment 215983 [details] [review]:

::: src/ephy-session.c
@@ -650,1 @@
 	   EphyEmbed *embed)

This is actually write_tab(), not ephy_session_save(), the comment is wrong, can you review that please?

@@ +665,3 @@
 	}
 	ret = xmlTextWriterWriteAttribute (writer, (xmlChar *) "url",
+					   (const xmlChar *) (fake_address ? fake_address : address));

Why not just freeing fake_address right here rather than at the end?

Review of attachment 215984 [details] [review]:

Seems about right!

Review of attachment 216099 [details] [review]:

::: src/bookmarks/ephy-bookmarks-editor.c
@@ +1473,3 @@
+    return result;
+}
+

To me this really looks like a deficience in the WK IconDatabase API, since there's no way to find out if a favicon exists for an url without allocating  a string. I think I'll file a bug in wk about this. Meanwhile I think this patch will do.

Review of attachment 216100 [details] [review]:

Another good catch.

Review of attachment 216100 [details] [review]:

Another good catch.

(In reply to comment #10)
> LRTM

What does this mean?

Created attachment 216134 [details] [review]
0005-ephy-session-fix-memory-leak-in-write_tab.patch

> the comment is wrong

Fixed

> Why not just freeing fake_address right here rather than at the end?

Thanks, I missed this.

(In reply to comment #18)
> (In reply to comment #10)
> > LRTM
> 

Looks right to me :)

Do you have a git account or should I push them for you?

(In reply to comment #20)
> Do you have a git account or should I push them for you?

I don't have git account, please commit.

I changed the variable name in the last patch, fake_address didn't sound very
good. Other than that, I pushed all the patches. Thank you!