798354 - Global variable sizecheck in ipc_channel_posix.cc is unused

Status: RESOLVED FIXED

(Core :: IPC, defect)

Comment 1

[(no longer active)]

Attachment: Patch (v1)

Comment 2

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Comment on attachment 668446 [details] [diff] [review]
Patch (v1)

r=me if it passes tryserver.

Comment 3

[(no longer active)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9adf5ca922a4

Comment 4

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/9adf5ca922a4

Comment 5

[(no longer active)]

Well, turns out that C++98 does not allow sizeof(class::member), but C++11 does.  Waldo tells me that it breaks builds using clang trunk.  Not sure what we can do besides backing out, so I did that:

https://hg.mozilla.org/integration/mozilla-inbound/rev/7bf353b1ae55

Comment 6

[[e-mail]]

I don't suppose making it extern would suffice to fool the compiler?

Comment 7

[(no longer active)]

(In reply to comment #6)
> I don't suppose making it extern would suffice to fool the compiler?

I don't know.

Comment 8

[John Schoenick [:johns]]

From a similar discussion on the clang-developers mailing list:

> sizeof(((sockaddr_un*)0)->sun_path)

Yeah...

Comment 9

[(no longer active)]

Neat trick!  Used that in https://hg.mozilla.org/integration/mozilla-inbound/rev/0761bc637081.

Comment 10

[Ed Morley [:emorley]]

(In reply to Ehsan Akhgari [:ehsan] from comment #5)
> Well, turns out that C++98 does not allow sizeof(class::member), but C++11
> does.  Waldo tells me that it breaks builds using clang trunk.  Not sure
> what we can do besides backing out, so I did that:
> 
> https://hg.mozilla.org/integration/mozilla-inbound/rev/7bf353b1ae55

https://hg.mozilla.org/mozilla-central/rev/7bf353b1ae55

Comment 11

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/0761bc637081