Hi Lex!
Le 16/08/2010 03:25, Lex Trotman a écrit :
Hi Colomban,
I am not criticising your attempt at improving this, instead I think the whole idea of trying to have the line and column info on the filename is wrong and will NEVER always be right until we make Geany a mind reader (although if its reading my mind it will probably still be wrong :-) because we can't think of all the ways "inventive" programmers will come up with to structure filenames with :nn on the end.
The +line --line and --column options are there to give an unambiguous position and the :line:col capability should be removed.
Even though I totally agree that the :line:column filename-suffix cannot be guessed right in all cases, and will never can, I think it is really convenient. Why? because many tools (compilers, grep, etc.) outputs the line/column information this way. This means that to open the correct file at the correct line after a `rgrep -n` the only thing I have to do is to copy the whole string (which generally can be selected in a single operation since it is considered as a single "word") and pass it as an argument to geany, and, how wonderful, it does exactly what I expect! The other syntax are good and ways stricter, but aren't as useful IMO because (I speak for me) we generally don't want to open a file on a specific line unless we got that line information from somewhere else -- which is often gcc or grep.
This said, again, I don't (want to) believe that people create /real/ files that ends with :nn. I played with this kind of suffixes only after seeing the bug report, and only for the sake of it. Then I'd totally understand that you -- or anybody reasonable, not me ;) -- don't want to waste more time on this "issue" :)
Just in case it was common I checked vim and emacs documentation and AFAICT neither does it. Emacs supports +line:col as an arg but not part of the filename.
I don't believe that this should prevent Geany to do so if it is a good idea :) But perhaps it means it isn't a good idea...
Cheers, Colomban