[SOLVED] vimscript: from current line only return list of every awk variable reference

porphyry5 · 09-07-2012, 10:07 AM

Working on an awk program in vim. From the current line only, I want to get a vimscript list of each awk variable or array reference in the line.
Here is an example of what I want to do done in bash.

Code:

~ $ grep -oP '[$a-z][a-z0-9\[\]]*[=, ;"/+*\-]' <<< "split(nt/3600,h,"."); m="." h[2]; split(m*60,mm,"."); s="."mm[2]; ss=s*60"|sed 's/.$//'|sort -u
h
h[2]
m
mm
mm[2]
nt
s
ss

grail · 09-07-2012, 11:38 AM

You may need to explain further as the example seems unclear to me? The line you have show is supposed to be what? A snippet of a bigger script? Also, you do not seem to have considered
any of the built in variables and your current example is easily broken when replacing the data (currently a period) between the quotes with characters such as letters.

Code:

$ grep -oP '[$a-z][a-z0-9\[\]]*[=, ;"/+*\-]' <<< "s="l"mm[2]" | sed 's/.$//'
s
$ grep -oP '[$a-z][a-z0-9\[\]]*[=, ;"/+*\-]' <<< "s="l"mm[2]; ss=s*60" |sed 's/.$//'
s
lmm[2]
ss
s

As you can see, both these outputs would appear to be incorrect.

porphyry5 · 09-07-2012, 01:22 PM

Quote:

Originally Posted by grail

You may need to explain further as the example seems unclear to me?

The awk scripts I've written have been simple and straightforward, and were no problem to get running. My latest one is more complex, and raises the question "how to debug awk?" --lint is not helping, so, after each line of awk code I want to add a line that prints the current value of all the variables referenced in the previous line.

By hand, that is going to be very tedious, so I want to write a vim script that will generate the desired print line from the content of the current line. Unfortunately, every time I try to write vimscript I enter an endless loop of help lookups to explain an unfamiliar term in the previous help lookup. After 4 or 5 of these I enter a comatose state in which I stare at the monitor for a while before erupting with "fu*k it, I'll write the damn thing in bash instead".

Which I have done in this case also. But it would be much cleaner if I did it within vim.

So here is the immediate problem. From a line of awk code, in vim extract each varaiable or array reference, and eliminate the duplicates.

ntubski · 09-07-2012, 03:34 PM

I haven't tried this myself, but GNU awk comes with an Awk Debugger.

porphyry5 · 09-07-2012, 04:10 PM

Quote:

Originally Posted by ntubski

I haven't tried this myself, but GNU awk comes with an Awk Debugger.

Thanks very much for this info. But in the current context, I want to get this vim scripting business off the ground, instead of just retreating to bash each time I try it.

I guess I'll mark this one as solved because I'm getting it done. I was unrealistic in expecting vim's pattern matching and extraction to be as sophisticated as grep's. Apparently you have to index your way through a string getting one match at a time, instead of all at once in a nice little array as with grep.