[tor-bugs] #25543 [Applications/Tor Browser]: Rebase Tor Browser patches for ESR60
Tor Bug Tracker & Wiki
blackhole at torproject.org
Mon May 14 15:08:15 UTC 2018
#25543: Rebase Tor Browser patches for ESR60
--------------------------------------+---------------------------------
Reporter: gk | Owner: arthuredelstein
Type: task | Status: needs_revision
Priority: Very High | Milestone:
Component: Applications/Tor Browser | Version:
Severity: Normal | Resolution:
Keywords: TorBrowserTeam201805 | Actual Points:
Parent ID: #25741 | Points:
Reviewer: | Sponsor:
--------------------------------------+---------------------------------
Changes (by gk):
* status: assigned => needs_revision
Comment:
Here is what I have so far. I am looking at 25543+8 and am moving upwards
the commits Arthur pointed out in comment:6. The full git commit hash is
the commit I looked at and the abbreviated in parentheses matches the
one(s) in comment:6.
The fixup commits make the reviewing a bit harder: Arthur, could you just
create new branches next time with stuff squashed (waiting until a
reasonably large amount of fixups have accumulated is fine, too). At least
having fixups on the review branch makes things harder for me. Anway, here
are my notes up to and including 4c9f746f2c19
15aab7d9640bebad5da4fe66535fe2646618cd73 -- good (b7ba24e9438c)
bc25026f2b7e77f5bc4b060fff7d52efc96a9286 -- good (3efb1fb5990a)
da7624af1f3d7cfd9878d1ef78f8212912bf00f3 -- good (7151b7736fbc)
7da50e7f9f892cb3f3e76d65e73f07051e771f02 -- good (83e40fc55843)
c078d465c98e224167c549ea03068e0d7455ad14 -- not okay (d9ffdac205cc)
fixed by 7e662c674e6089364aef5db63a24a1a78ea071ee
423b01125604d1ddcf5f0ddd7a613a9bfeb1969a -- good (7190f7e52771)
ef922fccd70ccf4def676097a652cc83d68684e7 -- not okay (6e18348d3fa2)
unresolved conflict
{{{
// Send tab key events.
+<<<<<<< HEAD
synthesizeKey("KEY_Tab", {shiftKey: activateShift});
+=======
+ var key = SpecialPowers.Ci.nsIDOMKeyEvent.DOM_VK_TAB;
+ utils.sendKeyEvent("keydown", key, 0, modifier);
+ utils.sendKeyEvent("keypress", key, 0, modifier);
+ utils.sendKeyEvent("keyup", key, 0, modifier);
+>>>>>>> febbeedade21... Bug 2874: Block Components.interfaces from
content
}}}
c567b8a14692dda0e43f01c9e472eb7b8244bb54 -- good (c91cc92acf64)
db22daac349cda4aeb7594451b1a37c4b73e0d5c -- good (05c64bde4a76)
3ac50c1adf35c931cbb89493b38d5a5c9d4662bc -- not okay (4fd7433d2b79)
not necessary as ripped out by the patch for 19910 (95ad1e098907), thus
both should go
a2acdd19161c87a6cb39f03ca03dde80c68f8e34 -- good (9a13c4dd4d89)
840833f13ee9cb64c8e2a248ffc43504dc828f09 -- not okay (f05b2599c291)
{{{
+ nsXPIDLString killTitle;
+ sb->FormatStringFromName(titleKey, params, 1,
getter_Copies(killTitle));
}}}
we should use |nsAutoString| as well and not |nsXPIDLString|
|killMessage| is now of type |nsAutoString|, so don't use
|getter_Copies()|
anymore like the unpatched code does
(seems this got fixed up in 0185a3889ea1f2d7e1262eeb5e3a4b04ba350d00,
though, which is good but confusing, could you fix this up in the commit
it belongs to?)
0185a3889ea1f2d7e1262eeb5e3a4b04ba350d00 -- okay (c8fbfdb5b0e7)
dbc9bd2f05b3728f222bc76b921542837124e237 -- okay (c38fc187252c)
1708d1f2330ee1766fd554841c3e5c054ec6baf2 -- okay (6bbe63c3f3b8)
8a0fca52fafbeaf92198d1daec00cbe1aaaffc59 -- okay (ba2620e0c91d)
1470089f3a2160f03e950536f1f0376b689f4b47 -- okay (73dc870c6712)
dad24d92f489ed935c2c79339f015c2ada6ccb4c -- okay (90f3c1b3b687)
ec4a23fc9ec9abbd2d6d7f8316ff9fc94bdc9950 -- okay (98966f5b88b5)
57f0a1e2a39e3b4431b3856f64c3877dc71f8dae -- okay (b4981a144854)
579734b9a25db60a7a9f59df198fe8cc2e7367e3 -- okay (d3a986dfb477)
a310a95c8dc23d152be27f404495c3e1ca036204 -- okay (ea9c5e94e364)
725b6c8f7d11010e922fe177b1cbda333a423720 -- okay (a71bf76df344)
a8add456af7aad04b0f4a402f1e5a8a7114aedec -- okay (5adf623b76f8)
[what about the other WebGL things we disable are we good with them wrt to
https://bugzilla.mozilla.org/show_bug.cgi?id=1217290?]
3d287f20252243b5c6cc41841fa2b48f90b3aa14 -- okay (aa65fd2ea82e)
6d25b3a1c946b70935f1df096878f0256292c143 -- okay (ac9bc3723c2b)
664d51393f8422a2b3d8edf5b24ee32b0a5d9b01 -- okay (90e817059ab7)
d93edac70a39e9f49555ae5cc96728b85448959a -- not okay (db5663390b3e)
fixed by cca5e99eb181e9e120a809e3ef71ea7e73fbec89
6dbef77faa36e1130612d0f6784b6b05c142dea8 -- okay (6a7ae76e406e)
do we still need it as the patch for 17009 got upstreamed?
81f89647957d420d087bbcd5c63104b157a139c9 -- okay (af9e23384692)
do we still need it as the patch for 15646 got upstreamed?
74ea0a667a7464d4a6e90cdd7a11ac60e06af649 -- okay (0a2323b8fcaa)
do we still need it as the patch for 1517 got upstreamed?
1261430503dd9a37a664e263b7415091cdf1f6a9 -- okay (fb26928c9f6f)
U 133a941a72c9 should be O 133a941a72c9 (in comment:6) as <isindex> got
removed
70c225a4b5f4266de7ce24110f5ef507160e234b -- not okay (b7f33de7c769)
there is no
dom/tests/mochitest/localstorage/test_localStorageByFirstParty.html
anymore, thus we can remove tbb-tests-ignore.txt it seems
c04c6fd4da01 is not a simple revert of db79c0270d50. The former only
reverts a part of it. That's due to a mess of this whole e10s compat
problem we needed to fix and me not properly dealing with the resulting
patches during rebase. So, we need the resulting patch after reverting
part of the first one for esr60 as
https://bugzilla.mozilla.org/show_bug.cgi?id=1305177 is still open.
856b810081cfbcd4b3f17ede4d1a95f4c02a14f0 -- not okay (d4da5714eb9d)
no need for those `extensions.hotfix` prefs
1cf891b3a783, 043e87d50499, and fdb2ad415cd6 are obsolete because we
decided we don't want to expose this to users anymore to make it harder
for them to shoot themselves in the foot?
fb7f56c28a9c3a77382361f218ac146c23ef6e30 -- okay (452829d9135f)
80e6743fbf4ceb60b1073e60efd6c67eb2a7fba0 -- okay (04e72287a8c7)
a761c12e657389847ff8c29ed836b5c2541c09c9 -- okay (87036e9e33eb)
The patch for 19411 is missing (fee72fffc081)
5c60f63e322184cfde30256b43ee7875ea2d1df6,
aede530c3ec8850224b7f6c49c1d7e1466684639 -- not okay (9ae35ba3c07e)
the first commit adds a .rej file;
I guess having now 'size: "1234"' is okay even though we had
it |null| before; however, I think we should stop patching the tests then
because there is no partial patch with an invalid hash anymore
I see:
{{{
- size: "1234",
+ hash: "1234cd43a1c77e30191c53a329a3f99d",
}}}
not sure how effective that is given |getLocalPatchString()|'s changed
signature
there is no SHA512_HASH_SIMPLE_MAR anymore:
https://hg.mozilla.org/mozilla-central/rev/e886bf1ba8fc removes it and we
don't backout that part (in aede530c3ec8850224b7f6c49c1d7e1466684639). So,
to sum up: just ignore the tests here I think
cdbdfbf912a00d38754e4da111647f89d0f7b70b,
c89b57cea0457ca7fcaa7338b23c57fbef38def2 -- okay (08964d93d418)
48e149a1f5f21aeda8f89d8dc374ede647feda23 -- okay (b0471f5e9e1f)
5c9aa89a778d10d19b55736abfcb63b745e14379 -- not okay (4c9f746f2c19)
{{{
- ACCEPTED_MAR_CHANNEL_IDS=firefox-mozilla-release
- MAR_CHANNEL_ID=firefox-mozilla-release
}}}
This should be "firefox-mozilla-esr" respectively (I guess you rebased the
patch before there was a dedicated esr60 branch)
aurora|alpha|beta|hardened|release|esr (and at other places) <- we don't
have the hardened series anymore
Do we need
{{{
- let prereleaseChannels = ["nightly", "aurora", "beta"];
+ let prereleaseChannels = ["nightly", "aurora", "alpha", "beta"];
}}}
?
The "extensions.lastTorBrowserVersion" part spilled over into
AddonsManager.jsm from #13052 + adding the AddonManager.jsm in
toolkit/mozapps/extensions/moz.build
Note to Arthur for the patch order: we need the patch for #13052 before
this one as we use GetTorBrowserUserDataDir()
{{{
+ done <"${tmpfile}"
}}}
add a whitespace between "<" and "\""
mcs/brade: how do you feel addressing #24476 while you are at it?
f8d84e457447ee34bdd4416191404f011fa4459e -- okay (506eb3cbd392 and
0f7641a6369c)
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/25543#comment:22>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
More information about the tor-bugs
mailing list