[Bioc-devel] heads up

Seth Falcon sfalcon at fhcrc.org
Tue Nov 28 06:43:49 CET 2006

Hi all,

I'm forwarding this for Duncan M.

> From: Duncan Murdoch <murdoch at stats.uwo.ca>
> Subject: Re: heads up
> To: Robert Gentleman <rgentlem at fhcrc.org>
> CC: bioc-devel at stat.math.ethz.ch
> Date: Mon Nov 27 13:42:36 2006 -0800
> Thanks for getting the bioc people involved.  I'm not currently
> subscribed to that list, so please cc me on discussions there, or post
> to the r-devel list.  (I'm not sure if this message will be allowed on
> the list, coming from a non-subscriber; if not, could you (Robert)
> please forward it?  Thanks.)
> Here is a quick summary of the current state:
> - Up to R 2.4.x, source(echo=T) and the equivalent code in Sweave()
> destroy formatting, and delete comments.  This is a bad thing.  See
> the Sweave FAQ 17.
>  - I've committed changes to R-devel (which will become 2.5.0
> eventually) that allow this problem to be fixed.  I'd like
> people to use them:  I think it's a good idea to format your code for
> readability, and to insert comments into it.
>  - The changes I committed are also designed to allow a lot of other
> things in the future, such as a source-based (rather than
> deparse-based) debugger.
>  - I understand that changing the way examples and vignettes look will
> cause short term trouble for people, and I'd like to minimize it, but
> without completely turning off the new behaviour (at least for now;
> I'll certainly review this decision before R-devel goes into alpha).
> I have put in place a scheme that means turning off the new behaviour
> is a one-line  modification to each vignette; I'd welcome suggestions
> for something that is even less work.
>  - The reason not to make the old behaviour the default now is that
> the new code needs to be exercised to find the bugs.  I don't want
> 2.5.0 to contain buggy versions when it is released.  Whether the old
> behaviour is made the default later will depend on input from users,
> how easy it is to turn it on and off, etc.
> Duncan Murdoch

More information about the Bioc-devel mailing list