The Humble Indie Bundle #2 Has Arrived

[m_gol]

Do you know why it was dysfunctional on 64-bits? It's because it wasn't a known problem until the testers moved to the problem configurations at the last moment. It's broken on 64-bit Intel with the latest libc- and pretty much only that. If you didn't have any testers testing with that, you'd have NEVER found it.

I've experienced the same problem with OpenArena on Ubuntu 10.04 64-bit (CPU: Intel Core2 Duo P8600), it didn't work because of strcpy used with overlapping strings, too. (by the way, it's still not fixed in Lucid repositories) So the issue would be present on the most popular desktop Linux distro in its 64-bit edition at least from April which gives us 8 months.

Nonetheless, I didn't want to say I don't appreciate your work. I don't regret money I spent on Humble2 even though I liked the first one more and there were less problems with it. I just think next time it would be nice to test all bundled games just for a couple of minutes each (if it installs, if menu works OK, if missions load properly etc.) about a month or two before deploying the bundle, at least on the most popular distros. If you (or whoever is responsible for the bundle) don't have enough manpower you could always give those games for free to just a couple of usual users in exchange for testing on their specific platforms. Just an idea.

[Xipeos]

Here we have Braid that doesn't work, Cortex Command providing disfunctional 64-bit version and Revenge of the Titans with a simple bug in deb creation

I do agree that the bundle was a bit rushed, but the devs are doing what they can to fix the issues. Unlike in the first bundle, most of these games were released on linux now for the first time (Braid, Cortex Command, RotT came out of beta), and release-day bugs were to be expected, as the indies don't have QA teams like the bigger companies. In HIB#1, the newest game (besides Aquaria) was more than a year old (World of Goo - february 2009).

The only unexcusable problems are Osmos's broken font, as they really should have fixed that for the bundle in the 1.5months since it was reported :-| (but there's an easy workaround at least) and Braid's lack of a proper readme (that maybe documents -width and -height so the game doesn't fail to start for people with small screens).

[Svartalf]

I've experienced the same problem with OpenArena on Ubuntu 10.04 64-bit (CPU: Intel Core2 Duo P8600), it didn't work because of strcpy used with overlapping strings, too. (by the way, it's still not fixed in Lucid repositories) So the issue would be present on the most popular desktop Linux distro in its 64-bit edition at least from April which gives us 8 months.

Heh... Not everybody was using the latest libc in my test pool and the bulk of the people that were, myself included, were using AMD in 64-bit mode (I've been doing X86_64 in 64-bit mode off and on since before ANYONE had it in hand except testers...in general, 64-bit works cleaner on AMD than Intel for high loads... ). Just because it's a "known" problem doesn't mean it would be found in the title because of testing. Do keep that one in mind (Though...I'd be a bit more than miffed at OpenArena there...that's a bit beyond the pale to know about it and not fix for that long. The moment I knew about it I was qualifying it even if it was too late to pull the plug on 64-bits and trying to find the fix for the past couple of days.... )

Nonetheless, I didn't want to say I don't appreciate your work. I don't regret money I spent on Humble2 even though I liked the first one more and there were less problems with it. I just think next time it would be nice to test all bundled games just for a couple of minutes each (if it installs, if menu works OK, if missions load properly etc.) about a month or two before deploying the bundle, at least on the most popular distros. If you (or whoever is responsible for the bundle) don't have enough manpower you could always give those games for free to just a couple of usual users in exchange for testing on their specific platforms. Just an idea.

I can't speak for the other teams, but...

1) They were tested on two of the popular distros for several months. They were hammered on in 64-bit mode extensively - but I didn't have any of the critical path configurations in hand in my teams or my systems and I was left scrambling to find out what the heck was going on right as the ship time happened. This isn't trying to cop-out; it's explaining just what went down so you can relate where I'm coming from and an apology for the something that almost always slips through the cracks at the last moment on a first release of something.

2) 32-bits works pretty well from what I understand...

3) The menu problem reported was a bit of a surprise- Mojoinstaller should have handled it right and nobody did the spaces there in my test pool. While it should have done it (I agree it should have- it's supposed to be one of the tools to be able to do that...), one wonders why one would put spaces in their directory path in the first place- many Linux applications don't like you doing that either... (Certain edge cases of command line utilities don't do well with the spaces and they choke. make being one of the ones that will get it wrong part of the time. I'm pretty sure that someone else would've fell upon the sword there. (Do keep in mind, I'm not going to comment either way on the other teams on this...

4) 64-bits is actually a bit of a new space for pretty much everyone. Even with my experience with it, I got zapped by a nasty surprise on ship.

[Svartalf]

I do agree that the bundle was a bit rushed, but the devs are doing what they can to fix the issues. Unlike in the first bundle, most of these games were released on linux now for the first time (Braid, Cortex Command, RotT came out of beta), and release-day bugs were to be expected, as the indies don't have QA teams like the bigger companies. In HIB#1, the newest game (besides Aquaria) was more than a year old (World of Goo - february 2009).

Thank you. I couldn't have described the circumstances or my thoughts on the matter better if I tried. I'm...heh...mortified at the two outstanding issues we've seen so far on the stuff.

[m_gol]

(Though...I'd be a bit more than miffed at OpenArena there...that's a bit beyond the pale to know about it and not fix for that long. The moment I knew about it I was qualifying it even if it was too late to pull the plug on 64-bits and trying to find the fix for the past couple of days....

It was fixed in OpenArena project, it just didn't end in Ubuntu 10.04 repositories. I provided a patch (acquired from the official version), but so far nobody seems to be willing to sponsor it. Here's the link:
https://bugs.launchpad.net/bugs/554748

I can't speak for the other teams, but...

1) They were tested on two of the popular distros for several months. They were hammered on in 64-bit mode extensively - but I didn't have any of the critical path configurations in hand in my teams or my systems and I was left scrambling to find out what the heck was going on right as the ship time happened. This isn't trying to cop-out; it's explaining just what went down so you can relate where I'm coming from and an apology for the something that almost always slips through the cracks at the last moment on a first release of something.

I guess it's just a price we pay for inhomogeneity of the Linux platform...

2) 32-bits works pretty well from what I understand...

Yep (except for Braid). I didn't try to install anything with a space in path, though. Also, I don't know if there were problems in that but it's in my habit to not install globally anything that didn't come in a native *.deb package. (it's not an option for multi-user setups, though).

3) The menu problem reported was a bit of a surprise- Mojoinstaller should have handled it right and nobody did the spaces there in my test pool. While it should have done it (I agree it should have- it's supposed to be one of the tools to be able to do that...), one wonders why one would put spaces in their directory path in the first place- many Linux applications don't like you doing that either...

Not everybody is required to know that. If it doesn't work it would be better to prevent it from the front-end side. But I understand you didn't know about the problem.

[Svartalf]

It was fixed in OpenArena project, it just didn't end in Ubuntu 10.04 repositories. I provided a patch (acquired from the official version), but so far nobody seems to be willing to sponsor it. Here's the link:
https://bugs.launchpad.net/bugs/554748

Useful info... Maybe it can get sorted out properly- that's not a good thing to have lurking in the mix. I carved it out wherever possible when I see it in code.

I guess it's just a price we pay for inhomogeneity of the Linux platform...

Indeed. But it's no different than some of the gotchas lurking in the Windows world (There's a lot most people blow off on other things that're more due to the kind of thing I got zapped with in that world...trust me... And it's actually a bit worse there than where we are... )

Yep (except for Braid).

Now, now... Be nice... It's as much S3's fault on that score.

I didn't try to install anything with a space in path, though. Also, I don't know if there were problems in that but it's in my habit to not install globally anything that didn't come in a native *.deb package. (it's not an option for multi-user setups, though).

I'm working on fixes for the games I've ported, but it's not a simple thing to just drop in PhysicsFS (though I've tried a hack that didn't work as well as I'd hoped...) though so it didn't happen with the rollout. As for multi-user setups...it depends on whether you allow them to locally install the game or not. If they do, each user can install.

Not everybody is required to know that. If it doesn't work it would be better to prevent it from the front-end side. But I understand you didn't know about the problem.

Believe me, I've got the same attitude as you do on that score...I'm working on sorting most of this stuff out and doing a re-spin pretty quick here.

[Svartalf]

As an update from the Twitter feed if you're not watching... They've pushed new RotT builds and they're propagating out to the content delivery network. Should be able to get un-busted packaging here shortly.

[m_gol]

As an update from the Twitter feed if you're not watching... They've pushed new RotT builds and they're propagating out to the content delivery network. Should be able to get un-busted packaging here shortly.

Will they be available via the official site?

[m_gol]

Now, now... Be nice... It's as much S3's fault on that score.

S3? Just for the record, my graphics card is Intel GMA 4500MHD.

[Remco]

S3? Just for the record, my graphics card is Intel GMA 4500MHD.

S3 protects Intel against lawsuits based on the S3TC patent.