Bug#407812: should check whether data is waiting instead of just reading

2007-12-11 Thread martin f krafft
also sprach martin f krafft [EMAIL PROTECTED] [2007.07.17.2029 +0200]: FWIW, Decklin, I can still see the problem with Debian sid's 0.13.0-1 The problem continues to be fully reproducible. I am now considering switching to xmms2, which works much better. If you need any assistance or testing,

Bug#407812: should check whether data is waiting instead of just reading

2007-12-11 Thread martin f krafft
Here is another interesting bit of information: I have two (HTTP) servers from which I stream to mpc/mpd. Both are behind the same firewall, which has three interfaces: 2: wlan: BROADCAST,MULTICAST,UP,1 mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:00:cb:02:2e:7b brd ff:ff:ff:ff:ff:ff

Bug#407812: should check whether data is waiting instead of just reading

2007-07-17 Thread martin f krafft
also sprach Decklin Foster [EMAIL PROTECTED] [2007.03.29.2035 +0200]: Nevertheless, I've stuck it in the SVN nightlies if anyone wants to experiment on it. They are running again in the usual place (http://musicpd.org/~decklin/debian). For Debian people: reports of success or failure with

Bug#407812: should check whether data is waiting instead of just reading

2007-03-29 Thread Decklin Foster
FYI: 0.12.2-2 has some older HTTP fixes from SVN merged, but I have not included the peek code because I can't fix the following issues: - Ogg over HTTP sometimes deadlocks again - All MP3s are reported with a length of 0 when playing; changing this breaks MP3s over HTTP in some way I don't

Bug#407812: should check whether data is waiting instead of just reading

2007-03-28 Thread Eric Wong
Decklin Foster [EMAIL PROTECTED] wrote: martin f krafft writes: retitle 407812 deadlock due to read() without select() Thanks, guys; my apologies for being MIA. I haven't gotten a handle on this yet, but I am able to reproduce it (consistently) without hitting any of the network code --

Bug#407812: should check whether data is waiting instead of just reading

2007-03-28 Thread Decklin Foster
Eric Wong writes: Warren made a patch that should fix the problem with HTTP below, OK, the inputStream_file.c does in fact fix the NFS case (at a very cursory test, anyway). For the HTTP case I will need to make a patch to pull in the changes to inputStream_httpRead in SVN before this one can

Bug#407812: should check whether data is waiting instead of just reading

2007-03-27 Thread Decklin Foster
martin f krafft writes: retitle 407812 deadlock due to read() without select() Thanks, guys; my apologies for being MIA. I haven't gotten a handle on this yet, but I am able to reproduce it (consistently) without hitting any of the network code -- my library on this installation is mounted via