Well, it seems my question got people stumped as much as I am! For an alternative solution, I would love advice on a (working) regex expression that would approximate the class of characters I am aiming at: say all the letters of the Latin1 set (ISO-8859-1). If we could eventually supplement that with Greek, Cyrillic, Arabic and Hebrew alphabets, that would probably cover more than all the ground I need.
2007/12/6, Pierre Thibaudeau <[EMAIL PROTECTED]>: > > I am using Struts 1 (more specifically 1.3.8). > > I have a form for which I use the Struts Validator. > > One of the textual fields should be able to accept any alphabetic > character, as well as spaces and apostrophes. > By "alphabetic characters", I don't merely mean [a-zA-Z], but I am also > interested in: > e acute, > a umlaut, > omega, > aleph, > and (while we're at it) Sanskrit characters. > But I want neither Arabic numerals, nor signs of punctuation (other than > the apostrophe), nor tabs, nor underscores, etc. > > I thought this might do it (but it doesn't work): > > <field property="name" depends="mask"> > <msg name="mask" key="error.nameMask"/> > <var><var-name>mask</var-name><var-value>^[\p{L}\' > ]*$</var-value></var> > </field> > > The symbol "\p{L}" (which means "any Unicode letter" in Java regex) > doesn't seem to do its job... > > Suggestion anyone? >