| 1 |
On Oct 14, 2009, at 5:36 PM, Michael Yang wrote: |
| 2 |
|
| 3 |
> hi guys, |
| 4 |
> |
| 5 |
> I'm on FreeBSD-7.2 x86 and have had to mask >portage-2.2.00.14200, |
| 6 |
> specifically: |
| 7 |
> |
| 8 |
> portage-2.2.00.14487 |
| 9 |
> portage-2.2.00.14555 |
| 10 |
> |
| 11 |
> This is due to emerge/ebuild hanging for a long time, usually with the |
| 12 |
> "checking XXX.tar.gz..." Running 'top' indicates that python2.6 is |
| 13 |
> stuck in ttywait, most likely in ebuild.sh. This runs for about 5 |
| 14 |
> minutes before a successful emerge proceeds. This does not happen on |
| 15 |
> Linux-x86 with the same portage versions. Since I'm no expert at |
| 16 |
> emerge/ebuild/portage, I was wondering if any one was noticing this |
| 17 |
> same behavior or could replicate it. |
| 18 |
> |
| 19 |
> -mike |
| 20 |
|
| 21 |
Hi, |
| 22 |
|
| 23 |
iirc according to truss it hangs in portage.__init__._test_pty_eof() |
| 24 |
when writing 'test_string = 2 * "blah blah blah\n"' to the terminal. |
| 25 |
If you make that method just return True the delay vanishes. Portage |
| 26 |
still seems to work then. |
| 27 |
|
| 28 |
-- Heiko |
Archives