Case insensitive variable substitutions

[pramsey]

Currently, variable substitutions are case-sensitive, and it would be useful to be able to handle substitutions case-insensitively at times. So, for example ​http://localhost/cgi-bin/mapserv?Subst=foo could match %subst% in the map file and substitute "foo".

The change will be mostly in msLayerSubstituteString, using strcasestr instead of strstr, when the feature is enabled. Handling of the _validation_pattern in the METADATA will also be changed to coerce the internal hash key to lower case before storing it and before testing against it.

FYI Note: This is a general purpose change, which will be put to specific use in passing the LAYERS key from a WMS request into a PostGIS data statement, that will split the WMS LAYERS string and use the values to pull information from a metadata table and write appropriate layer copyright information into the acetate map overlay.

Enabling case insensitive substitutions will be done with a CONFIG directive:

CONFIG SUBSTITUTIONS_IGNORE_CASE [yes|no]

The existing case-senstive behavior will be the default.

[pramsey]

While working on the patch, I found that the layer->metadata hashtables are already case-insensitive. The only case-sensitive piece left is the actual substitution in msLayerSubstituteString. I'm wondering if making the case-insensitivity the default behavior might be OK?

[sdlime]

I was kinda wondering the same thing too, but even broader. The substitution code from mapstring.c is used primarily in 3 places:

• the runtime subs
• templates subs
• logical expression subs

There's already a case insensitive version of msReplaceSubstring() available in mapstring.c (as msCaseReplaceSubstring()). I don't know why we don't use that instead of the case sensitive version everywhere- unless there are big performance penalties. I wouldn't think that would break apps.

This isn't a RFC-able enhancement although running the msCaseReplaceSubstring() for msReplaceSubstring() by mapserver-devs would be a good idea.

Steve

[pramsey]

And I was wondering if we could replace msCaseFindSubstring() throughout the code base with strcasestr()!

[pramsey]

I've placed the larger issue in #3254, so I can deal with my small issue here. Do you think the %substitution% code in particular can be safely made case insensitive?

[sdlime]

Yes. Anyone relying on %foo% to behave differently from %FOO% is asking for trouble regardless. I think in 6.0 we simply go case insensitive for tags or all forms...

Steve

[pramsey]

Committed at r9672

[pramsey]

Closing as complete!