On Sunday 19 April 2009 23:55:14 Kai Blin wrote:
As mentioned by the authors of Publish or Perish in their FAQ (http://www.harzing.com/pop_faq.htm#Q309), Wine exporting sxs.dll and not providing a real CreateAssemblyCache function trips up the installer. This patch allows Publish or Perish to install without issues.
Thanks to Donnie Berkholz for pointing me at this issue.
dlls/sxs/sxs.c | 10 ++++++++++ dlls/sxs/sxs.spec | 2 +- 2 files changed, 11 insertions(+), 1 deletions(-)
Of course the moment I sent this, I found the real implementation in fusion.dll. Sending a forward, feel free to ignore this patch.
Cheers, Kai
2009/4/19 Kai Blin kai.blin@gmail.com:
On Sunday 19 April 2009 23:55:14 Kai Blin wrote:
As mentioned by the authors of Publish or Perish in their FAQ (http://www.harzing.com/pop_faq.htm#Q309), Wine exporting sxs.dll and not providing a real CreateAssemblyCache function trips up the installer. This patch allows Publish or Perish to install without issues.
Thanks to Donnie Berkholz for pointing me at this issue.
dlls/sxs/sxs.c | 10 ++++++++++ dlls/sxs/sxs.spec | 2 +- 2 files changed, 11 insertions(+), 1 deletions(-)
Of course the moment I sent this, I found the real implementation in fusion.dll. Sending a forward, feel free to ignore this patch.
Are other methods in sxs.dll forwardable to fusion?
- Reece
On Sun, Apr 19, 2009 at 5:04 PM, Kai Blin kai.blin@gmail.com wrote:
On Sunday 19 April 2009 23:55:14 Kai Blin wrote:
As mentioned by the authors of Publish or Perish in their FAQ (http://www.harzing.com/pop_faq.htm#Q309), Wine exporting sxs.dll and not providing a real CreateAssemblyCache function trips up the installer. This patch allows Publish or Perish to install without issues.
Thanks to Donnie Berkholz for pointing me at this issue.
dlls/sxs/sxs.c | 10 ++++++++++ dlls/sxs/sxs.spec | 2 +- 2 files changed, 11 insertions(+), 1 deletions(-)
Of course the moment I sent this, I found the real implementation in fusion.dll. Sending a forward, feel free to ignore this patch.
I suspect this is the incorrect approach. Anastasius Focht writes in bug 15877 (http://bugs.winehq.org/show_bug.cgi?id=15877#c3):
"SxS (aka win32 fusion) is different from fusion for CLR assemblies. Although both Sxs and CLR have the concept of GAC (for Sxs it's WinSxs) there are various technological differences."
Thus it may not be entirely reasonable to forward the sxs.dll functions to fusion.dll. I provisionally attempted the same for the LiteStep 0.3 installer in bug 17631 and found that while forwarding sidesteps the unimplemented stub crash, it didn't allow the LiteStep installer to properly detect the Visual C++ 2005 runtime.
-
Andrew Nguyen -
Kai Blin -
Reece Dunn