Dear Linux users,

I have a sed script that I use to automatically correct many of my most common writing mistakes in LaTeX documents. This sed script contains more than thousand lines in the form:
(...)
s/betwenn/between/g;
s/bilatteral/bilateral/g;
s/bussiness/business/g;
(...)

I apply this script to my .tex files to correct all mistakes and overwrite my original .tex file. But, sometimes I would like the script to ask me to confirm every change, that is, for every matching regexp I would like the script to ask me if I want to proceed with the modification. I am sure there is an option for that, but I have not been able to find it yet. Thus, I would greatly appreciate any help or hints to solve this problem. Many thanks,

Regards,

Gilles

Hi.

The solution is probably to not use sed. 'aspell' has a tex mode, so it can spell check your tex source files without tripping over the markup language, and it can work interactively.

Dave

Hi,

thanks for your reply ilikejam.

In fact my script performs more advanced modifications than just correcting simple spelling mistakes (which I could clearly do with any spell checker). For example,

# replace (...)^ by {(...)}^;
s/\(([^)(]*)\)\^/{\1}\^/g;
# replace max by \max in equations (only if they are delimited by $ symbols);
s/\(\$[^\$]*[^\\]\)max\([^i][^\$]*\$\)/\1\\max\2/g;

I could translate my script to another language though, but I'd like to keep a list of regular expressions separate from the program files, so that it is easier for non-programmers to add new regexp and above all that the list of regexp is easily reusable. Any suggestion about the language to use to do that ?

Gilles

You could, I suppose, run each sed command seperately, outputting to a different file, then run diff on the original and the new file, and ask if the changes are OK. Pretty trivial to script, but it's gonna be slow if you've got a lot of commands.

You might be able to do something a bit more elegant with awk, though.

Dave