Recommended builds anyone?

[mhebsacker]

Hi,

on my iHP 140 I still run the build from 18.07.06 after I found it to be pretty stable as long as Tagcache isn't used.

I tried newer builds two or three times but always was unlucky to immediately run into issues (f.e. two months ago the audio buffer was simply a mess) which made me go back.

After I am no developer but happy user of Rockbox I dont have much of an interest to download new builds every few days just to figure out they do not work (usually while sitting in a train with no chance to revert to a different version). So - all I want is a stable build! Which is the reason I still stick with the old build from July.

This was discussed a few times already but after it seems that working towards a release is none of the top priorities (at least this is my impression) I'd ask again if it would'nt be possible to at least introduce something like "recommended builds" which have proofen to be more or less trouble free (lets say "usable builds", not release quality).

Would be a great help for all who simply want to use Rockbox, but dont have the time or technical skills to deal with troubled daily builds. With just daily builds labeled as "we have no idea if this build actually might work" Rockbox is for sure shiying away many users - or loosing those who tried Rockbox, but by accident downloaded one of those builds with more serious issues.

Regards;

Matthias

[pabouk]

I install the latest CVS builds almost every day and I have not run into many serious problems during the last year and a half. AFAIK the todays build runs pretty well. I recommend it

[LinusN]

The thing is that we want as many users as possible to run the latest versions, otherwise we don't get the feedback we need.

[L]

Pretty sure most daily builds are stable enough to use. I think the trouble you're facing comes from not updating the builds properly.

[Llorean]

If everyone downloads specific older builds, nobody is providing feedback, and some bugs are so rare that only large groups of users have a statistical chance of finding them.

Remember: Until there is a version numbered release, there are no users, only testers. If you consider yourself a user, that's your choice, but the software doesn't *have* a user version yet, so all you're doing is being a bad tester. Wholly your choice, mind you, but not aligned with actually improving the project.

[mhebsacker]

Yes, bad tester might be, but what if I'm very happy features whise since quite a while, am very happy to have a far better replacement to the original iRiver Firmware (gapless playback beeing the most serious issue), and all I'm looking forward to is to get rid of the last bugs of what is already on my player?

My motivation to try a new build is very low as long as I always have the risk to just download new bugs instead of having old ones fixed, even if the new bugs are accompanied by new features I don't need.  I simply want to listen to music and not observe another different strange behaviour of my device.

So I'd very motivated to provide feedback and test every other build in a feature freeze phase, but somehow completly lost interest to download and test builds with new bits added every day after I figured out that after the features freeze was liftet problems I'd with the new builds where increasing instead of decreasing (and yes, I deleted preferences a.s.o.) which was when I reverted to an old build and stayed with it.  This was a completly different picture than in the first half of this year, where I'd the impression that new builds had a tendenci to work better.

This probably is completly different for developers, I understand that improoving a software feature wise is far more interesting than bug hunting. If the developers are happy to develop Rockbox only for themselves and are not interested in users it's there right to do so.

Regards;

Matthias

[Llorean]

If you provide accurate feedback on how to reproduce bugs, you can help get them fixed. If you download a new version, see that it has new bugs (regardless of whether or not it's fixed the old ones) and go back to your old build, those bugs may never get fixed simply because they may not affect everyone else (or better still, too many people chose to ignore them and roll back instead of actually caring about helping fix them.)

There are a lot of developers who do strive to fix bugs, the problem is there are VERY few good bug reports. During a feature freeze, the ONLY bug reports are about old bugs. During this period, most people are either doing what you're doing (avoiding new bugs by holding back) or simply updating to a newer version every day and hoping the bug goes away, rather than filing a proper report with instructions to reproduce. If you look at the bug tracker, very few filings contain useful information. And in the last 14 days, there have been a whole 4 bugs filed.

Your complaint seems to be about the lack of bugs being fixed, but many bugs aren't experienced by everyone, and developers CANNOT fix bugs they don't know about. With people refusing to take the chance that there will be new bugs in their builds, or not bothering to report them (you seem to have claimed to encountered several new bugs, how many of them have bug reports filed explaining how to reproduce them?), how are developers expected to know that these bugs are there to be fixed?