[PATCH 1 of 6] changegroup: expose list of changesets from getsubsetraw

Pierre-Yves David pierre-yves.david at ens-lyon.org
Mon May 25 23:59:49 CDT 2015

On 05/25/2015 09:53 PM, Gregory Szorc wrote:
> On Mon, May 25, 2015 at 9:32 PM, Pierre-Yves David
> <pierre-yves.david at ens-lyon.org <mailto:pierre-yves.david at ens-lyon.org>>
> wrote:
>     On 05/25/2015 05:42 PM, Gregory Szorc wrote:
>         # HG changeset patch
>         # User Gregory Szorc <gregory.szorc at gmail.com
>         <mailto:gregory.szorc at gmail.com>>
>         # Date 1432589936 25200
>         #      Mon May 25 14:38:56 2015 -0700
>         # Node ID ea28b7239085bbc7d561ab6aa6f86bdc19f77a25
>         # Parent  605b1d32c1c011d56233f28923ee5354fce7e426
>         changegroup: expose list of changesets from getsubsetraw
>         Currently, when generating changegroups, it is possible for the
>         caller
>         to not know exactly what data is inside without parsing the returned
>         data stream. This is inefficient and adds complexity.
>         Return the list of changesets from getsubsetraw to make this data
>         more widely available.
>     I'm confused. This very data is -provided- to the function. It is in
>     `outgoing.missing`. Why would we need to return it?
> Err, I may have gotten too carried away with looking at return values
> and a handful of functions that were named very similarly. I'll send a v2.

What function remains that are not taking outgoing. In the bundle2 
context I expect about all of them to be in the modern form and take an 
outgoing object.

If it still make sense to modify some of them (And I'm not use it does , 
since old function we mostly left in place for old caller) you probably 
want to return the outgoing object that get computed internally.

Pierre-Yves David

More information about the Mercurial-devel mailing list