[mapserver-dev] Enhancement to RFC-40 conditional wrapping

[Stephen Woodbridge]

Daniel Morissette wrote:
> In RFC-40, we added the ability to do conditional wrapping, that is 
> wrapping labels only if they exceed a specified length:
> 
> http://mapserver.org/development/rfc/ms-rfc-40#line-wrapping
> 
> One of our users reported a problem with labels such as 
> "Notre-Dame-du-Rosaire" or "Saint-Pierre-de-la-Rivière-du-Sud"... I 
> didn't make up those names, municipality names in Quebec all use 
> hyphens, and some are very long and can sound quite "interesting".
> 
> Anyway, the problem is that when we use conditional wrapping on those 
> labels, using the hyphen as the wrap character, we end up with something 
> like:
> 
>   Notre-Dame
>   du-Rosaire
> 
> we lose the second hyphen. What we would like to get is:
> 
>   Notre-Dame-
>   du-Rosaire
> 
> Thinking about this some more, I realized that in the case of 
> hard-wrapping (maxlength < 0) we may want the ability to insert a hyphen 
> at the end of each line. It's not that simple in French, but in English 
> one could wrap "someveryverylongword" and expect:
> 
>   someveryve-
>   rylongword
> 
> Perhaps the solution to both problems would be to add the ability to 
> specify a character to insert at the end of each line? Not sure what to 
> call it... WRAPTEXT, WRAPEOL, WRAPMARK, ???
> 
> Or another more scalable option could be to add a WRAPMODE param, with 
> the different modes being
>  - wrap on each instance of the WRAP char
>  - conditional wrap on maxlength
>  - hard wrap on maxlength
>  - and for each case above, insert or not a hyphen (or other char) at 
> end of line, or keep the wrap char there instead of removing it from the 
> string

Daniel,

I like the idea of adding WRAPMARK <char> to specify and control whether 
you mark the wrap point with <char>.

I also think that adding WRAPMODE <mode> or LABELMODE <mode> would make 
the behavior much more explicit and obvious, which I would hope would 
make it easier for people to understand and use. Having to lookup or 
memorize the behaviors in the proposed matrix of len == 0|<0|>0 and 
existance of WRAP is not very obvious.

Also it is not obvious that no WRAP and MAXLENGTH >0 should skip the 
label or truncate the text of the label to MAXLENGTH! Should 
TRUNCATE|SKIP be explicit modes?

<mode>
EACH
CONDITIONAL
HARD
SKIP
TRUNCATE

or something like that.

-Steve W