Bug 1578

Summary: SSTATE_MIRROR needs a trailing '/' character to work
Product: [Build System, Metadata & Runtime] Meta-yocto Reporter: Matthew McClintock <msm-oss>
Component: meta-yoctoAssignee: Richard Purdie <richard.purdie>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Medium CC: msm-oss, poky.bs.watcher, poky.watcher, sgw
Version: unspecified   
Target Milestone: 1.2   
Hardware: x86   
OS: Multiple   
Whiteboard: Final patch out for review on bitbake-devel
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: ---

Description Matthew McClintock 2011-10-03 13:25:10 UTC 
If I set 

SSTATE_MIRROR = "http://linux.freescale.net/yocto/sstate-cache"

the sstate mirror will not be used, however if I simply change this too:


SSTATE_MIRROR = "http://linux.freescale.net/yocto/sstate-cache/"

it works again.
Comment 1 Richard Purdie 2011-11-04 10:22:03 UTC 
I'm a bit confused by this report - there is not a SSTATE_MIRROR variable and SSTATE_MIRRORS doesn't use that syntax? Do you mean:

file://.* http://linux.freescale.net/yocto/sstate-cache

doesn't work but:

file://.* http://linux.freescale.net/yocto/sstate-cache/

does? That is a natural part of the mirror mapping code unfortunately...
Comment 2 Matthew McClintock 2011-11-04 12:06:01 UTC 
(In reply to comment #1)
> I'm a bit confused by this report - there is not a SSTATE_MIRROR variable and
> SSTATE_MIRRORS doesn't use that syntax? Do you mean:
> 
> file://.* http://linux.freescale.net/yocto/sstate-cache
> 
> doesn't work but:
> 
> file://.* http://linux.freescale.net/yocto/sstate-cache/
> 
> does? That is a natural part of the mirror mapping code unfortunately...

Correct. 

But its extremely non-obvious when you miss this. As we are just looking for cached stuff and not finding the cache is not a fatal result. Therefore it's easy to just assume something else is broken with the caching system.

Maybe we can add a QA on this variable to check the a trailing '/'? Or maybe it's better to leave it alone. Not sure what the best approach is...