Hi Dag! On 23.08.05, Dag Wieers wrote:
On Tue, 23 Aug 2005, Joerg Lehmann wrote:
Please test extensively as there are many, partially rather invasive, changes included in this release. In particular, a new database version requires an update. You, thus, should backup your song databases before running the new release for the first time and ensure that there is enough disk space for the upgrade.
Hi,
First report:
+ At first the error was not clear that I had to convert my database from:
~/.pytone/music2a.db -> ~/.pytone/music2a/db
only when checking what pytone did to have an empty database I noticed the dbname/db file. I love this change though, makes configuration much easier.
Yes, this was a mess in the first place. I hope the conversion doesn't lead to too many problems, though.
+ 450MB free diskspace was clearly not enough to convert a 7MB database from v4 to v5. And it only did 2500 records out of 3500. After making room, it used 820MB for 3440 records, and it seemed to slow down the longer it takes and diskspace usage is not directly related to number of records. (It seems to require progressively more)
+ The 920MB was not enough for my biggest database, it got stuck at 3500 records out of 3799 and consumed the complete 920MB free space. I had to free up 1.1GB for a 7MB database with only 3799 records.
Sounds familiar to me. Maybe, we should checkpoint the database in between. I didn't do that because during the whole upgrade, we're in a single transaction (in order to avoid breaking the whole db, when something goes wrong during the upgrade). Ok, I just tried that and it didn't help :-(
+ After doing the last database, I noticed it said: 4 years, no clue what that means :)
Not too much, just that you had 4 different years in your database. Didn't there appear a "Done" afterwards?
+ Also it did not play automatically, just like before. The only way to make it play music is by starting it with -d test (and in return have a debug file that grows and grows :))
Don't know how to debug this though :/
Have you tried -d /dev/null. Note, however, that debugging leads to a noticeable performance degradation. Otherwise, the only thing I can suggest is to add log.info calls in the player code in order to find out where it hangs.
PS Changing the speed is so funny :) It would be nice to have a window showing how much it is increased (much like the volume actually).
Yes, I also noticed that. And I also would suggest that we make the change "multiplicative" and not "additive" as it currently is. This corresponds much more to the way human perception works. Jörg