On Sat, 07 Mar 2020 00:39:17 -0500 Daniel Kahn Gillmor <d...@fifthhorseman.net>
wrote:
> Package: src:rust-mach-o-sys
> Version: 0.1.1-2
> Control: affects -1 src:rust-memsec
> 
> https://buildd.debian.org/status/package.php?p=rust-mach-o-sys
> 
> shows that many debian platforms are failing to build.
> 
> The failures mostly look like this:
> 
> ```
> 581 | pub const BIND_SPECIAL_DYLIB_FLAT_LOOKUP: ::std::os::raw::c_char = -2;
>     |                                                                    ^^
> cannot apply unary operator `-` |
>     = note: unsigned values cannot be negated
> ```
> 
> I think they're happening on platforms where char is unsigned by
> default.
> 
> Given that mach-o-sys is really intended for the OS X platform, it isn't
> surprising that it's failing.
> 
> However, this causes memsec to be BD-uninstallable on those same
> architectures.
> 
> The only reason i can see for having rust-mach-o-sys in debian is for
> the ability to cross-compile to MacOS.  But I've never tried to do that,
> and maybe it's not actually possible.
> 
> Anyway, i'm tempted to carve it out from rust-memsec, since it's only
> relevant there:
> 
>   https://codesearch.debian.net/search?q=mach.o.sys&literal=0&perpkg=1
> 
> If we do carve it out of rust-memsec, we might want to just remove
> rust-mach-o-sys from debian until someone has a concrete need for it.

4 years passed and no one cared enough for it. Plus upstream never touched it
since 2016. Even the then only rdep memsec moved away. Time for a RM?

-- 
Sdrager,
Blair Noctis

Attachment: pgpSBev73b8z4.pgp
Description: OpenPGP digital signature

Reply via email to