A solution might be that the GWT editor outputs {{xhtml}}character{{/ xhtml}} This should always work. WDYT? -Vincent On Sep 24, 2008, at 2:31 PM, Vincent Massol wrote:

Hi,

We have a problem with the GWT character map. It contains some characters that are not ISO-8859-1. The problem is that since our default encoding is ISO-8859-1 (<?xml version="1.0" encoding="ISO-8859-1" ?>) the browser replaces for example this character ∞ with ∞

Our rendering doesn't accept HTML so this value is escaped and is shown as is.

So we have several non-satisfactory options:

1) Switch to UTF8 by default. A good thing but that won't solve the problem for those using the GWT editor in ISO-8859-1 encoding. 2) Modify the GWT editor character map so that it only shows characters valid with the current encoding. This is a pity since it should be possible to display other chars since the browser knows how to display them even though the user is not in the correct encoding 3) Introduce a wiki syntax for representing characters not allowed in the current encoding. However this means we'll still need to replace ∞ with this syntax before the rendering is called and this forbids users to enter ∞ in the wiki editor so this is not good.

Any better idea?

WDYT?

Thanks -Vincent

Erm, is there any valid reason for any XWiki installation not to change encoding when upgrading version? The only reason I see is clients that have set "for this URL force this encoding". I really think 1) is the solution and will be the only one to approach IRIs in some future. So the config should say explicitly that using iso-xxx would break some parts of the GWT-WYSIWYG editor (it's really just not iso-8859-1 alone!). paul Le 24-sept.-08 à 14:31, Vincent Massol a écrit :

Hi,

We have a problem with the GWT character map. It contains some characters that are not ISO-8859-1. The problem is that since our default encoding is ISO-8859-1 (<?xml version="1.0" encoding="ISO-8859-1" ?>) the browser replaces for example this character ∞ with ∞

Our rendering doesn't accept HTML so this value is escaped and is shown as is.

So we have several non-satisfactory options:

1) Switch to UTF8 by default. A good thing but that won't solve the problem for those using the GWT editor in ISO-8859-1 encoding. 2) Modify the GWT editor character map so that it only shows characters valid with the current encoding. This is a pity since it should be possible to display other chars since the browser knows how to display them even though the user is not in the correct encoding 3) Introduce a wiki syntax for representing characters not allowed in the current encoding. However this means we'll still need to replace ∞ with this syntax before the rendering is called and this forbids users to enter ∞ in the wiki editor so this is not good.

Any better idea?

WDYT?

Thanks -Vincent

_______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs