[SOLVED] Excluding pattern-matched characters from replacement in sed

[calebt3]

I need to correct frequently-occurring typos in several html files using sed.

A common one is my search pattern being interrupted by newlines. Google told me that I could use

sed ':a;N;$!ba;s/\n/ /g'

to turn newlines into spaces, but this will leave messy html. What I want to do is remove every newline not preceded by a closing angle bracket.
I tried

sed ':a;N;$!ba;s/[^>]\n/ /g'

But that will remove the last letter/character as well as the newline.
Is there any way I can keep that letter from getting chopped out?

[calebt3]

After hours of searching, I finally break down and as the question myself. Of course I find the solution on my own immediately after.

Answer: wrap the part I want to keep in parenthesis (in my cast, the non-bracket character): \([^>]\) and add /1 to the replacement text in the place I want it.

sed ':a;N;$!ba;s/\([^>]\)\n/\1 /g'