1431801 - Sandbox MinGW Compilation errors: Don't declare duplicate instantiations

Status: RESOLVED FIXED

(Core :: Security: Process Sandboxing, enhancement)

Comment 1

[Tom Ritter [:tjr] (OOTO until 4/30 at least)]

> 0:14.80 /home/tom/Documents/moz/mingw-work/sandbox/security/sandbox/chromium/base/strings/utf_string_conversion_utils.cc:125:71: error: duplicate explicit instantiation of ‘void base::PrepareForUTF8Output(const CHAR*, size_t, std::__cxx11::string*) [with CHAR = wchar_t; size_t = unsigned int; std::__cxx11::string = std::__cxx11::basic_string<char>]’ [-fpermissive]
> 0:14.80  template void PrepareForUTF8Output(const char16*, size_t, std::string*);
> 0:14.80                                                                        ^
> 0:14.80 /home/tom/Documents/moz/mingw-work/sandbox/security/sandbox/chromium/base/strings/utf_string_conversion_utils.cc:146:71: error: duplicate explicit instantiation of ‘void base::PrepareForUTF16Or32Output(const char*, size_t, STRING*) [with STRING = std::__cxx11::basic_string<wchar_t>; size_t = unsigned int]’ [-fpermissive]
> 0:14.80  template void PrepareForUTF16Or32Output(const char*, size_t, string16*);
> 0:14.81                                                                        ^

Comment 2

[Tom Ritter [:tjr] (OOTO until 4/30 at least)]

Attachment: Bug 1431801 Don't declare duplicate instantiations.

On MinGW, these typedefs are the same, and mingw complains about duplicate instantiations.
Rather than use -fpermissive, just comment out the second instantiation.

Review commit: https://reviewboard.mozilla.org/r/214360/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/214360/

Comment 3

[Tom Ritter [:tjr] (OOTO until 4/30 at least)]

Comment on attachment 8944035 [details]
Bug 1431801 Don't declare duplicate instantiations.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/214360/diff/1-2/

Comment 4

[Bob Owen (:bobowen)]

Comment on attachment 8944035 [details]
Bug 1431801 Don't declare duplicate instantiations.

https://reviewboard.mozilla.org/r/214360/#review220070

r+ - I suppose, but is this also a MinGW bug?

Comment 5

[Tom Ritter [:tjr] (OOTO until 4/30 at least)]

(In reply to Bob Owen (:bobowen) from comment #4)
> r+ - I suppose, but is this also a MinGW bug?

I think it's a compiler oddity, and not really a bug.  The types are the exact same:
https://chromium.googlesource.com/chromium/chromium/+/master/base/string16.h#37

gcc complains, but would allow it if we switch on -fpermissive. msvc (and therefore clang-cl) apparently don't complain.

Comment 6

[Bob Owen (:bobowen)]

Ah yeah sorry I misread the first one.

Comment 7

[Tom Ritter [:tjr] (OOTO until 4/30 at least)]

Patch land order: 1431621, 1431801 (this bug), 1431807, 1431825

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a132c57bc190
Don't declare duplicate instantiations. r=bobowen

Comment 9

[Dorel Luca [:dluca]]

https://hg.mozilla.org/mozilla-central/rev/a132c57bc190