Re: [Bug-indent] C++ ability for indent .. code patch and question.

[Preston A. Elder]

Comments inline:

Carlo Wood wrote:

> Aren't there people using a format like:

> if (a>b) {
>  a=b;
> }
> though?

>

> If they are, and they start with:
>
> if (a > b) {
>  a = b;
> }
>
> then the resulting code would be:
>
> if (a >b) {
>  a=b;
> }

AFIK, theres no option in indent to 'compress space around operators', 
tho I could be wrong.  It will ALWAYS format the above as:
if (a > b) {
never as:
if (a>b) {
no matter what options or input you give it, so I dont see the problem 
there.  This is, infact desirable behavior, since indent is a pretty 
printer, not an ugly printer ;)

> I think that the correct strategy is to never catenate two '>' tokens.
> If might be better to add a flag to parser_state_tos that indicates
> that the last token was a '<' and then request a space before the
> current '<' token if there was also a space before it in the input.


You are correct, more to the point, the correct strategy would be to 
never concatinate 2 operators (of any kind, not just '<', '>', etc). 
The only exception I suppose would be '!' (!!strcmp(blah,blah) is valid, 
bad and ugly, but valid).  I did some testing before I made the change, 
if I put ANYTHING between the operators (be it a comment, some text, 
pretty much anything really), they would not concatinate together -- so 
that might help you in you're search for cecation of operator 
concatination ... I just dont know the code anywhere NEAR well enough to 
make that kind of change -- my change is indeed, a hack, and I'll be the 
first to admit it, however, it works :)


Anyway, do with it as you will,

--
PreZ
Development Head
Magick IRC Services

Magick IRC Services, power to the PEOPLE!
Product Site: http://www.magick.tm
Project Site: http://www.sourceforge.net/projects/magick