code review 13607045: os.user: Fix C.malloc to use C.ulong instead of C.size_t.

os.user: Fix C.malloc to use C.ulong instead of C.size_t.

[dirk, 2013-09-07 04:37:56]

Hello golang-dev@googlegroups.com, I'd like you to review this change to https://code.google.com/p/go/

[dave_cheney.net, 2013-09-07 05:14:24]

On 2013/09/07 04:37:56, dirk wrote: > Hello mailto:golang-dev@googlegroups.com, > > I'd like you to review this change to > https://code.google.com/p/go/ Thanks for taking a crack at this, i'm guessing this is addressing issue 5926. If so please include the magic works, Fixes issue 5926. in the issue description. While you're there, please update the description to be the more standard, os/user: fix C.malloc to use C.ulong instead of C.size_t This change will probably need lots of interop testing.

[dave_cheney.net, 2013-09-07 05:14:25]

On 2013/09/07 04:37:56, dirk wrote: > Hello mailto:golang-dev@googlegroups.com, > > I'd like you to review this change to > https://code.google.com/p/go/ Thanks for taking a crack at this, i'm guessing this is addressing issue 5926. If so please include the magic works, Fixes issue 5926. in the issue description. While you're there, please update the description to be the more standard, os/user: fix C.malloc to use C.ulong instead of C.size_t This change will probably need lots of interop testing.

[minux1, 2013-09-07 13:19:44]

i think this will compile in certain toolchain configurations, but not all. 

[iant, 2013-09-08 04:34:54]

NOT LGTM Unfortunately it's not that simple. With this change the GNU/Linux (Ubuntu Precise) build fails with # os/user lookup_unix.go:64: cannot use _Ctype_ulong(bufSize) (type C.ulong) as type C.size_t in function argument

[rsc, 2013-09-09 14:40:15]

I am hoping that this will be fixed before the official Mavericks release, not that I have any reason to believe it will. I asked Andrey Mirtchovski to run Go at tip on a Mavericks pre-release a few days ago and he reported that all.bash passed. Perhaps you and he are using different versions of Xcode? What does 'xcodebuild -version' print for you? Thanks. Russ 

[aam, 2013-09-09 14:59:00]

> What does 'xcodebuild -version' print for you? the previous log I sent you was with: $ xcodebuild -version Xcode 4.6.3 Build version 4H1503 i'll update to the latest "developer preview 7", but while that'd downloading, attached is a log with that exhibits the error: $ xcodebuild -version Xcode 5.0 Build version 5A11365x 

[rsc, 2013-09-23 17:32:46]

The new cgo should accept C.size_t correctly, even with the buggy Xcodes. 

[saysjonathan, 2013-09-30 14:14:11]

> I asked Andrey Mirtchovski to run Go at tip on a Mavericks pre-release a > few days ago and he reported that all.bash passed. Perhaps you and he are > using different versions of Xcode? What does 'xcodebuild -version' print > for you? I'm also seeing build failures on Mavericks pre-release with both Xcode 5 DP5 and GCC-4.2: ; xcodebuild -version Xcode 5.0 Build version 5A11365x ; gcc-4.2 --version i686-apple-darwin11-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3) Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. GCC version was tested by setting CC=/path/to/gcc-4.2. ; echo $CC /usr/local/bin/gcc-4.2 ;./all.bash <snip> # os/user pkg/os/user/lookup_unix.go:64: cannot use _Ctype_ulong(bufSize) (type C.ulong) as type C.size_t in function argument

[rsc, 2013-12-18 04:21:06]

R=close (We fixed this a different way.)