Re: problem with gtk-docs being corrupt

[Malcolm Tredinnick <malcolm commsecure com au>]

Brian,

On Wed, Dec 18, 2002 at 05:04:33PM +0000, Brian Cameron wrote:
> I am not on the gtk-doc-list, so please cc: me with any response.
> 
> Funny characters are appearing in the docs that I am generating 
> via gtk-docs.  Generally the pages don't look so nice.  I'm including
> an attachment as an example.
> 
> An ASCII example which shows the non-Latin characters showing up
> in the page is below...
> 
> Any tips regarding how to make the pages look nicer would be 
> appreciated.

OK, we've seen this before in a couple of different scenarios. My guess
would be that you are using a slightly older version of gtk-doc and the
pages are not being correctly interpreted as UTF-8 encoding.

A potted history: The most recent gtk-doc code encodes the pages in
ISO-8859-1 (changed on November 16). Prior to that, it was outputting
UTF-8 and, for a short period, not correctly setting the charset value
in appropriate meta tag it produced. However, even after setting the
charset correctly, we were seeing problems because Apache, by default,
was serving the pages as ISO-8859-1. There are changes that can be made
to Apache's config file to make it server those pages as UTF-8, but in
the interests of making life easier for everybody, James changed the
pages back to ISO-8859-1.

So the solution is to upgrade gtk-doc to something recent (e.g. the 0.10
release or the CVS version) and go from there. If you are _already_
using the latest and greatest version, can you let us know how you are
viewing your pages (via a web server? Just directly from files?) and
what encoding your browser thinks they are, etc.

Hope this helps.

Cheers,
Malcolm

-- 
The only substitute for good manners is fast reflexes.