Uprev libmojo to r456626 |
|||||||||
Issue descriptionhis adds support for typemapping and is a pre-requisite of changing the Mojo protocol in a way that we can have protocol negotiation.
,
Jul 18 2017
ping. Any progress here?
,
Sep 12 2017
Ping. Libchrome was patched to r456626 almost 6 months ago on Mar 22, but the libchrome ebuild still hasn't been uprev'ed yet. This blocks several efforts to make changes to libchrome, see e.g. crbug.com/737950 and crbug.com/751899. There's a work-around involving patching the ebuild (see CL:497434), but it's less than ideal. Is there any progress on this? If the project is currently on ice, can c3f34a3eed92e804017c0eaf9a9c69fd2f39d2ec be temporarily reverted until it is continued?
,
Sep 12 2017
Sorry for the delay. We are blocked because the AOSP libmojo patch was reverted on AOSP, because it broke the Mac build (and it needs to land before we can land the rest of the patches). Turned out the AOSP Mac bots use Yosemite and our box was a Sierra (the failure only happens on Yosemite). We should be getting a Yosemite box any day now and resumes the effort very shortly.
,
Sep 13 2017
Thanks for the update. I realized I probably won't have to change libchrome right now, so I won't have to do the patch work (pun!) again and I'm glad libchrome gets an uprev at all, so thanks for doing this!
,
Nov 23 2017
Sorry to nag, but Libchrome is still in an inconsistent state. The code you see in the local checkout isn't what Chrome OS actually uses, which can be very confusing. It took me a bit today to realize why JoinString doesn't accept a vector<StringPiece> (r456626 does, but r395517 doesn't). Are there any updates when this will be resolved? Thanks!
,
Mar 23 2018
Friendly ping - is there any plan to continue this uprev? The current "half-done" situation forces any changes to be put into the EBUILD's files/ directory, which is weird. It's especially quirky for the stuff that was actually present *both* in r395517 and in r456626, but was deleted from libchrome in between (in commit [1]) and therefore isn't present in the 7b88bc885b9d8dc551beab840b853a79fa06494d libchrome revision used by the ebuild. [1] https://chromium.googlesource.com/aosp/platform/external/libchrome/+/f6024733c0d1eed88f68520b5e6a20b96e212ad6
,
Mar 23 2018
Reassigning this bug to lhchavez@ who knows more about the state of the uprev.
,
Mar 26 2018
+Eric, who has added a lot of libchrome patches lately, preparing the uprev. +Hidehiko, who wrote a design doc ob libchrome uprevs: https://docs.google.com/document/d/17uw7GxK3oNIK0pLuRKupg2Wcd272wl4GIi2ms8SDjXo
,
Oct 16
,
Nov 12
,
Nov 12
Any update? This is blocking an important security update for the JSON parser in libchrome. The current state with r456626 checked into libchrome, but not used in Chrome OS, makes the CL unreviewable since it has to be submitted as a patch. https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/1125732 https://bugs.chromium.org/p/chromium/issues/detail?id=860181
,
Nov 12
Re #12. Here are CLs. CL:1239818, CL:1240053, CL:1239834 and CL:1240033 Those are LGTM'ed once, but CQ found a failure on some specific boards. The fixes are in CQ already for several days: CL:1323529, CL:1323589, CL:*693349. Though, we had lab outage, and it does not seem to up yet, which blocks them...
,
Nov 28
|
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by djkurtz@chromium.org
, Jun 30 2017