Re: [talk] why it'd be great to have web interface for submittingsimple doc patches

[Phil Holmes]

----- Original Message ----- 
From: "Joseph Rushton Wakeling" <address@hidden>
To: "Phil Holmes" <address@hidden>
Cc: "James" <address@hidden>; <address@hidden>
Sent: Saturday, October 06, 2012 4:32 PM
Subject: Re: [talk] why it'd be great to have web interface for 
submittingsimple doc patches


> On 10/06/2012 05:21 PM, Phil Holmes wrote:
> > Unfortunately, testing that docs compile cleanly takes about 15 times as 
> > long as
> > code, so it's not for the underpowered or faint hearted.  Used to be 2 
> > 3/4 hours
> > on my virtual machine.
>
> Yes, true.  The from-scratch build time for docs is pretty hefty, though 
> from my memory once you've got it built _once_, building an edited version 
> takes much less time.  The build system improvements that came in a while 
> ago (am I right this was your work?) made it much better than it was a few 
> years back.


Most of the tidying up was me.  There were a number of other improvements in 
the way dependencies were calculated that Julien was responsible for.  John 
M has also made some improvements.

As you say, compile-edit-compile cycles are shorter than the full build, but 
can occasionally not reveal errors, so for a proper test it's always better 
to nuke the build directory and rebuild from scratch.

And given the amount of effort involved in all this, using git-cl seems a 
small additional step.

--
Phil Holmes