• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: architecture dependant problems with ResMerger
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: architecture dependant problems with ResMerger


  • Subject: Re: architecture dependant problems with ResMerger
  • From: Eric Albert <email@hidden>
  • Date: Fri, 13 Apr 2007 00:44:43 -0700

It's been a long time since I worked on this, but here's what I remember....

-ResMerger does use resource flippers when building on Intel. That's because it has to read the resource from each file and then write out the result, and whenever we read or write resources on Intel we flip them. Technically it wouldn't have to do this because the resource data itself is just a binary blob as far as merging resource files is concerned, but that's just the way the code happened to work.

-Custom resource types naturally can't be flipped here, but as described above it isn't really a concern.

-Flippers aren't supported on PowerPC, so there's no way to test this there.

-The error message is probably valid. It's telling you that the WIND flipper believes that your resource is shorter than a WIND resource is supposed to be. If that WIND resource works on the system anyway, it's a bug in the flipper. This happens from time to time -- some parts of the system assume that if a resource is shorter than expected the last fields are 0 and use that instead of reporting an error. More often, though, it's a bug in the resource, and loading that resource actually doesn't work.

Hope this helps,
Eric

On Apr 12, 2007, at 6:08 PM, Rua Haszard Morris wrote:

Interesting - it seems the intel build of ResMerger calls the resource flippers or at least checks resource sizes. Does it do this with custom resource types? It won't have access to any custom flippers. Hmm.

Sounds like a good bug report for bugreporter.apple.com.

I just spend a few hours finding out that ResMerger behaves differently on PPC than on i386, more specifically it does give errors on Intel that it does not on PowerPC:

One of our resources fails when building on Intel with this error:
### /Developer/Tools/ResMerger ERROR: errCoreEndianDataTooShortForFormat (resType 'WIND' ID 5)


Building the same project on PPC does not throw any errors whatsoever. The resulting app will run fine on PPC Macs, but fail the same error message on Intel.

I could find a post in the mailing list archives where this is being explained with the byte swapping mechanism: http:// lists.apple.com/archives/xcode-users/2006/Sep/msg00344.html

Is there any way to make the PPC ResMerger complain about that error too? Our main build machine is a PowerMac G5, and it would be very beneficial if it could detect such errors.

Thanks,
Stefan

_______________________________________________ Do not post admin requests to the list. They will be ignored. Xcode-users mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: This email sent to email@hidden
References: 
 >architecture dependant problems with ResMerger (From: Stefan Werner <email@hidden>)
 >Re: architecture dependant problems with ResMerger (From: Rua Haszard Morris <email@hidden>)

  • Prev by Date: Re: architecture dependant problems with ResMerger
  • Next by Date: Setting environmental variables for xcode executables
  • Previous by thread: Re: architecture dependant problems with ResMerger
  • Next by thread: Re: architecture dependant problems with ResMerger
  • Index(es):
    • Date
    • Thread