I'd like to go ahead and ask for
./run_indent.pl `egrep '*\.[ch]$' MANIFEST | egrep -v \
'global_setup.c|packfile.c|mops.c|io.h'`
to be applied to CVS.
I've noted some glitches in indent's output, summarized below. However, I
think it's easier to fix these afterwards than for me to try to mail a giant
patch to the list.
Here are some of the anomalies i've noticed.
- indent seemed to not like the "const" in (encoding,chartype)_lookup,
and tried to put the return value on the same line.
- it also put the return value on the same line for string_chopn.
- it tried to do "||strcmp" at one point- i added the space back in.
- need to reformat a few contents (replace * ** with *)
- it adds an extra space around function call derefs like this:
(*run_native) ().
- it did this: if ((interpreter->flags & PARROT_PREDEREF_FLAG)!=0)
(note no spaces around the !=)
- it introduced spaces before the parens on calls to
mem_sys_memcpy
string_*
This is due to the fact that my run_indent.pl program needed to pick
up #defined types. So it's picking up macros which aren't really
types, i'm afraid. That'll need a better solution, clearly.
Notice that i'm omitting a few files which need more extensive tweaking:
- mops.c has fancy comments which shouldn't be messed with- leave
it alone for now.
- io.h gets ugly.
- it did something very odd to the switch cases in packfile.c.
- global_setup.c got real ugly.
Yes, this will cause headaches and giant diffs (approx 11000 lines) but if not
now, when? On the plus side, it cuts down the errors from "make check_source"
by 162..
--Josh
--
Josh Wilmes ([EMAIL PROTECTED]) | http://www.hitchhiker.org
