| Summary: | net/nicotine does not start | ||
|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Tim J. Robbins <tim> |
| Component: | Individual Port(s) | Assignee: | Tilman Keskinoz <arved> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | ||
| Priority: | Normal | ||
| Version: | Latest | ||
| Hardware: | Any | ||
| OS: | Any | ||
|
Description
Tim J. Robbins
2004-03-29 10:40:15 UTC
State Changed From-To: open->analyzed Yes, this looks amd64 specific. I copied my working config files from my i386 and I get the same error. Responsible Changed
From-To: freebsd-ports-bugs->perky
Something is wrong with the bsddb python module.
perky, can you look at this?
% cat bdbtest
import bsddb
x = bsddb.hashopen("bsddb.test")
x["one"] = "1"
print x
x.close()
x = bsddb.hashopen("bsddb.test")
print x
% python bdbtest
Traceback (most recent call last):
File "bdbtest", line 2, in ?
x = bsddb.hashopen("bsddb.test")
File "/usr/local/lib/python2.3/bsddb/__init__.py", line 192, in hashopen
d.open(file, db.DB_HASH, flags, mode)
bsddb._db.DBInvalidArgError: (22, 'Invalid argument -- architecture lacks fast mutexes: applications cannot be threaded')
On Mon, Mar 29, 2004 at 07:52:17AM -0800, Tilman Linneweh wrote:
[snip]
> File "/usr/local/lib/python2.3/bsddb/__init__.py", line 192, in hashopen
> d.open(file, db.DB_HASH, flags, mode)
> bsddb._db.DBInvalidArgError: (22, 'Invalid argument -- architecture lacks fast mutexes: applications cannot be threaded')
>
The origin of this problem is db41's mutex detection.
On i386:
% grep db_cv_mutex build_unix/config.log
db_cv_mutex=x86/gcc-assembly
% grep HAVE_MUTEX_THREADS build_unix/db_config.h
#define HAVE_MUTEX_THREADS 1
But, on amd64:
% grep db_cv_mutex build_unix/config.log
db_cv_mutex=UNIX/fcntl
% grep HAVE_MUTEX_THREADS build_unix/db_config.h
/* #undef HAVE_MUTEX_THREADS */
So we can't let DB_ENV_THREADS flag on and Python should be built
with WITHOUT_THREADS=yes to support bsddb enough on amd64.
Or, fixing db41 port would be better than that. :)
Hye-Shik
Responsible Changed From-To: perky->arved Grab back, as perky is right, it is not a python but a db41 isssue. I am working with the maintainer on a fix. State Changed From-To: analyzed->feedback Adding --with-mutex=x86/gcc-assembly for db41 seems to fix this problem. I am currently waiting for freedback from the maintainer regarding the possible impact. State Changed From-To: feedback->closed Please update to the latest db41 and try again. nicotine seems to work here now. |