Re: E patches for review: SQL update object and fix for seedVat envExtras

Kevin Reid <kpreid-M/[email protected]> Tue, 12 Oct 2010 15:28:37 -0400

Newsgroups	gmane.comp.lang.e.general
Message-ID	<[email protected]>

On Oct 11, 2010, at 9:10, Thomas Leonard wrote:

> Two small patches, both to code I added:
>
>
> 1. Return an object from SQL updates, not simply an int
>
> http://gitorious.org/~tal-itinnov/repo-roscidus/it-innovation/commit/877d5e55e203c7c42515e229b163e87b2c7297eb
>
> This allows the caller to fetch any generated keys returned by the
> update.

This seems fine to me.

> 2. Use eParser to unserialize data-E in seedVatAuthor
>
> http://gitorious.org/~tal-itinnov/repo-roscidus/it-innovation/commit/cf42d22410cfcd61219ce0e3e0b8f520840084ee

Since the string was known to be produced by a surgeon using the  
standard exits, this is OK. Please add a comment to that effect.

> This is faster,

Yep; it removes the interpretation overhead of the Data-E  
implementation. This should hopefully improve in fast E  
implementations (compilers, cheaper dispatch).

> and avoids a bug handling cycles. See:
>
> http://www.eros-os.org/pipermail/e-lang/2010-October/013624.html

I'd forgotten about looking into that; I'll try to get to it.

-- 
Kevin Reid                                  <http://switchb.org/kpreid/>