Re: [PATCH 2/4] MIPS/GAS: Reformat the SVR4 PIC JAL test

["Maciej W. Rozycki" <macro at codesourcery dot com>]

On Wed, 21 Jul 2010, Richard Sandiford wrote:

> > Given the function is fully annotated with .ent, .end, etc. I 
> > couldn't resist the temptation to complete symbol attributes by adding a 
> > .size directive.
> 
> Have you noticed any practical difference after adding .size?  .ent and
> .end should set the size for you, so I'd prefer to keep .size out.

 Oh, I'm surprised as I didn't recall it being the case -- though 
obviously I may have relied on behaviour I observed some ten years ago 
that may have been changed/fixed since (or I could have simply got 
confused at some point and misremembered).  Dropping the change then.

> In general, I think it's a bad idea to make cosmetic changes like adding
> a space after commas.  That's very much down to personal preference;
> some asm coders prefer the space whereas others don't.  I don't think
> adding a space after each comma really qualifies as a "correction".
> But go with what you've got as far as that's concerned.

 I wouldn't do anything like this if the source wasn't ill-formatted in 
the first place, which suggests whoever wrote it could have no style at 
all in the first place.

 In this particular case if I kept no space after commas, I would have to 
use:

	.size	text_label,.-text_label

or worse yet:

	.size	text_label,. - text_label

(as the GNU coding standard asks for spaces around binary operators) for 
consistency, so the choice was obvious (to me anyway) and I updated the 
rest of code accordingly.

 Since I've dropped .size I can drop these changes too.  Thanks for the 
review.

  Maciej