| 1 |
On Tuesday, June 28, 2011 12:37:41 justin wrote: |
| 2 |
> On 6/28/11 6:23 PM, Mike Frysinger wrote: |
| 3 |
> > On Tuesday, June 28, 2011 02:54:03 Michał Górny wrote: |
| 4 |
> >> I think that also a good idea may be to provide an Makefile example, |
| 5 |
> >> showing that often sed is unnecessary, and it's enough to do things |
| 6 |
> >> like: |
| 7 |
> >> |
| 8 |
> >> emake CC="$(tc-getCC)" CFLAGS="${CFLAGS}"... |
| 9 |
> > |
| 10 |
> > this is easily dangerous when it comes to packages (and many do) that |
| 11 |
> > append in the Makefile. specifying on the command line blocks those |
| 12 |
> > while passing via env works fine. i'm not sure it's appropriate to |
| 13 |
> > provide as an example. -mike |
| 14 |
> |
| 15 |
> There are cases and cases. You are right with what you said. But cases |
| 16 |
> where the makefile appends stuff are not the typical cases where a sed |
| 17 |
> is applied. Normally sed is used if upstream hardcodes FLAGS we don't |
| 18 |
> want. |
| 19 |
|
| 20 |
and those packages should get patched and the patch sent upstream which means |
| 21 |
the sed+cmdline is still unnecessary. |
| 22 |
-mike |
Archives