[Sparkle] A Question of Enclosures

[Adam Radestock]

On 6 Apr 2008, at 3:55 pm, Rob Napier wrote:

> On Sun, Apr 6, 2008 at 8:43 AM, Adam Radestock <raddish at glassmonkey.co.uk 
> > wrote:
> SparkleCaster currently does not do differential binary patches. I  
> know this is mentioned in the original post that Andy made about  
> SparkleCaster, but I have not implemented it yet. Is this something  
> that
>
> a) would be feasible (given the code-signing issues, among others),
>
> I think it should be as long as part of the binary diff includes the  
> new signature. As long as the final signature matches the final  
> bundle, it shouldn't matter how it got that way.

That's true; after re-reading the documentation on code signing, you  
can use it with binary patches.

- For those of you who have not read about the new code signing  
feature in 10.5, the main advantage for developers (especially ones  
that use Sparkle), is that your users will not have to see that  
annoying "this application has changed. Allow it to access your  
keycahin?" message, as the OS knows that it's already a trusted app.

>
>
> b) developers would actually use?
>
> Oh yes. One of my project's .app bundle is over 70MB (and over 100MB  
> when you include all localizations). The vast majority of that is  
> resources that seldom change. I would very much like to push  
> incremental updates. The majority of our users will move from  
> release n-1 to n (and we're using Sparkle to help ensure that).
>

Ok, I'm having a look into MojoPatch now to see if I can use it to  
generate a diff for SparkleCaster.
This causes me to revisit the previous point of storing the enclosures  
locally after the appcast has been uploaded. Presumably, we'd need to  
cache at least the n-1'th enclosure to run the diff comparison on?

On a different note, I think that I might archive this disscussion on  
my Trac wiki, if no-one objects? It might be useful for us to refer  
back to in the future...

Kind Regards,

Adam Radestock
Glass Monkey Software
www.glassmonkey.co.uk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.andymatuschak.org/pipermail/sparkle-andymatuschak.org/attachments/20080406/349c4acc/attachment.htm