Wed Feb 22 09:04:04 2012: Request 75193 was acted upon.
Transaction: Correspondence added by steve...@planit.com
Queue: PAR-Packer
Subject: RE: [rt.cpan.org #75193] [PATCH] Fix build on Win32/VC++
Broken in: 1.012
Severity: Important
Owner: RSCHUPP
Requestors: s...@cpan.org
Status: patched
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75193 >
Roderich Schupp via RT wrote on 2012-02-21:
> Tue Feb 21 10:10:31 2012: Request 75193 was acted upon.
> Transaction: Correspondence added by RSCHUPP
> Queue: PAR-Packer
> Subject: [PATCH] Fix build on Win32/VC++
> Broken in: 1.012
> Severity: Important
> Owner: RSCHUPP
> Requestors: s...@cpan.org
> Status: open
> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75193 >
>
> On 2012-02-21 09:32:44, SHAY wrote:
>> The current version (1.012) of PAR-Packer does not build on Win32 with
>> VC++. There are two problems: myldr/boot.c attempts to include
>> VC++unistd.h
>
> Thanks for the patch, these changes were already in trunk.
>
Oops, sorry. I should have checked there first.
>> myldr\mktmpdir.c(45) : warning C4013: 'getuid' undefined; assuming
>> extern returning int
>
> I wrapped the whole function calling getuid (isSafeDir) in #ifndef
> WIN32 as it's only called on *nix anyway.
>
> OTOH if someone could implement isSafeDir for Windows, I surely would
> appreciate that :) It's there to fix a CVE "PAR::Packer creates per-
> user directory with unsafe permissions".
> I'm not sure what it takes to create a "safe per-user directory"
> on Windows, i.e. where only the user (and privileged accounts) can
> create and modify files.
>
> If you want to test the current snapshot (esp. since you use the VC++
> toolchain), there were some recent changes in the resource handling for
> Win32 executables
>
> http://svn.openfoundry.org/par/PAR-Packer/trunk
>
Works fine for me, except that 20-pp.t still fails test 32 as before. Is that
supposed to be fixed now? Do I need a newer Win32-Exe than I've got? I'm using
0.17.
(The error is different now: it now complains about not being a valid Win32
executable, rather than complaining that its side-by-side configuration is
incorret.)
