| 1 |
Whats wrong with: |
| 2 |
emerge pkgname &>/some/place/repfile.rep |
| 3 |
? |
| 4 |
It should be easy to call it inside the script you use to emerge packages on |
| 5 |
this server. If you desire to make reports on a per-package basis it should |
| 6 |
be easy to create the script that would run emerge --pretend packagename |
| 7 |
first and then call emerge for every reported package and capture reports |
| 8 |
separately... |
| 9 |
WTF incorporating this into portage itself: I did not see this mentioned |
| 10 |
(though there were some related things suggested). If you think this is an |
| 11 |
important addition please file a feature request to bugs.gentoo.org. |
| 12 |
|
| 13 |
George |
| 14 |
|
| 15 |
|
| 16 |
On Sunday 09 June 2002 17:27, Rufiao Valhacouto wrote: |
| 17 |
> In the process of creating a build server (a server which is capable of |
| 18 |
> building all the packages in the portage tree to test if a given profile |
| 19 |
> is broken, and maybe providing binary packages for people who desire |
| 20 |
> them) I've found it was required to have portage to be able to log the |
| 21 |
> [per package] ebuild output into a file (more likely in the /var/db/pkg/ |
| 22 |
> tree), so that it could be analyzed later. In fact, this behavior is |
| 23 |
> desirable generically, since it's sometimes useful to understand what |
| 24 |
> happened in the building process of a given package. Is this |
| 25 |
> functionality being planned for portage? Do you see any drawbacks to |
| 26 |
> include it in the mainstream code? |
Archives