On Thu, 4 Nov 2010, Eric Noulard wrote:
> 2010/11/4 Michael Stefaniuc <[email protected]>:
> > Julia Lawall wrote:
> >> On Thu, 4 Nov 2010, Michael Stefaniuc wrote:
> >>
> >>> Hello,
> >>>
> >>> the SmPL code
> >>> @@ @@
> >>> + lock();
> >>> bar( ... );
> >>> + unlock();
> >>>
> >>> applied to
> >>> void foo(void)
> >>> {
> >>> bar(a,
> >>> b);
> >>> }
> >>>
> >>> will produce following diff
> >>> + lock();
> >>> bar(a,
> >>> b);
> >>> + unlock();
> >>>
> >>> The indentation level of the code inserted after the multi-line function
> >>> call matches that of the last argument and not that of the function
> >>> identifier. I have tested this with versions 0.2.3 and 0.2.4rc5.
> >>
> >> Thanks for the report. I'm not surprised that it does this, but I'm not
> >> sure that I can easily fix it either. It just has a stream of tokens when
> >> it is doing the unparsing, not a parse tree. It is also thrown off when
> >> you add a statement after a conditional where the branch is indented.
> >> That case is indeed probably harder, because there is a complete statement
> >> on the line before.
> >>
> >> Anyway, I will look into the function call case.
> > Thanks but don't worry too much about this as it isn't a real problem.
> > In most cases the diff needs to be edited anyway to fix the "every
> > argument on a separate line" style.
> >
>
> Couldn't your patched file be postprocessed with indent or astyle
> in order to enforce certain coding style?
>
> Or may be coccinelle itself may call such tool only on patched file?
> ...or even only the portion of code which was patched, but I don't
> how indent/astyle would behave on "code snippet" (never tried).
It could. But at least for Linux code, I wouldn't want to change the
indentation in the unmodified part.
julia
_______________________________________________
Cocci mailing list
[email protected]
http://lists.diku.dk/mailman/listinfo/cocci
(Web access from inside DIKUs LAN only)
