Review: plugin touchups

While writing some long-overdue documentation, I noticed a lot of small improvements that could be made to some of the format plugins.

* Several places incorrectly set "ResolutionUnits" attribute that should have been "ResolutionUnit".

* A few places incorrectly set "BitsPerSample" attribute to the bits per pixel.

* In icooutput.cpp, use boost::iequals rather than #ifdef and stricmp/strcasecmp.

* In some other places, use iequals to do case-insensitive comparisons of attribute names and values (and prefer it over strcmp).

* Added additional file extensions to JPEG and Targa.

* A few other miscellaneous attribute names or values were regularized to conform to our usual conventions.

* Softimage input neglected to set the ImageDescription to the comment field.

[Chris Foster, 2011-01-08 08:40:12]

Mostly looks good, a few comments below

http://codereview.appspot.com/3832044/diff/1/src/bmp.imageio/bmpoutput.cpp
File src/bmp.imageio/bmpoutput.cpp (right):

http://codereview.appspot.com/3832044/diff/1/src/bmp.imageio/bmpoutput.cpp#ne...
src/bmp.imageio/bmpoutput.cpp:161: res_units.compare ("m") == 0) {
not sure why compare() is being used here instead of operator==() :

res_units == "pixels per meter" || res_units == "m"

http://codereview.appspot.com/3832044/diff/1/src/softimage.imageio/softimagei...
File src/softimage.imageio/softimageinput.cpp (right):

http://codereview.appspot.com/3832044/diff/1/src/softimage.imageio/softimagei...
src/softimage.imageio/softimageinput.cpp:169: m_spec.attribute
("ImageDescription", comment);
So... the reason you're not copying directly out of m_pic_header.comment into
the spec is that the final character of the 80 char buffer may be non-null and
you want to preserve it?

In that case the code seems to need comment[80]=0 to be fully safe.

also, why not

if(m_pic_header.comment[0] != 0) {
    // do the copy etc.
}

[larrygritz, 2011-01-08 16:19:27]

Thanks, Chris.  I think you are correct on both counts, and I will adjust it
before commit.