Tim's recent post got me looking at \z
I've read :h :syn-ext-match but it's gibberish to me at the moment
the \zs and \ze modifiers are handy ways ot indicating where any
replacements start/end. Their documentation can be found at
:help /\zs
:help /\ze
The
:help syn-ext-match
is a red-herring, as you end up in a very different section of
the docs if you just do
:help \z
as you discovered :)
If you want to delete everything after the 2nd comma, you can use
:%s/,[^,]*,\zs.*/
I get the 'search the whole document for a comma followed by a not-
comma followed by whatever then a comma'
then ???
In the above, the ",[^,]*," (which you have correctly
understood/described) specifies context. The "\zs" means "when
performing the replacement, start the replacement *here*, rather
than replacing the entire match thus far". This would be
behaviorly identical to
:%s/\(,[^,]*,\).*/\1
which preserves the "comma, not-a-comma stuff, comma" bit of
things rather than deleting them.
Hope this sheds more light than it obscures,
-tim
