| 1 |
* Nicolas Kaiser [nikai@×××××.net] [2003-09-06 09:42 +0200]: |
| 2 |
|
| 3 |
> > for i in `find /usr/portage/${my_target_subdir} -name "*.ebuild" -print` ; |
| 4 |
> > do egrep -l "^KEYWORDS=\".*~ppc.*\"" $i >| output.txt ; |
| 5 |
> > done |
| 6 |
> |
| 7 |
> Ah, but you could also increase speed vastly by not using a loop: |
| 8 |
> (with both of them reading from memory) |
| 9 |
> |
| 10 |
> $ time for [loop] |
| 11 |
> real 1m8.892s |
| 12 |
> user 0m31.380s |
| 13 |
> sys 0m41.600s |
| 14 |
> |
| 15 |
> $ time find [xargs] |
| 16 |
> |
| 17 |
> real 0m1.053s |
| 18 |
> user 0m0.810s |
| 19 |
> sys 0m0.900s |
| 20 |
|
| 21 |
Your probably right but I wouldn't say `time` is a good benchmark, you |
| 22 |
will have the results of the egrep cashed from the first run .. you |
| 23 |
cheat ;) |
| 24 |
|
| 25 |
Anyhow .. I took sometime today to update the script. You can now search |
| 26 |
for -arch and I attempted to impliment no-arch (that is the abscence of |
| 27 |
the search_string from the ebuilds) .. It doesn't work :P and I'm not |
| 28 |
quite sure, nor do I have time right now to figure out, why exactly. |
| 29 |
Prehaps someone else has a clue? It's not alltogher neccessary but would |
| 30 |
round the script off and so any pointers welcome (also the sed could |
| 31 |
probably be less verbose). |
| 32 |
|
| 33 |
best |
| 34 |
|
| 35 |
cal |
Archives