For your gaming pleasure, we now present Onlink v0.2.0-a8, over 300 commits since v0.2.0-a7, and a lot of changes, too! But before anything, here are the appropriate download links:

Windows | Mac | Linux

Windows users: don't forget that all alpha builds will be under the 'Beta' distribution for now. Again, this isn't the final v0.2.0 release, and we expect quite a bit more work before we get to that version.

Mac users: we're pretty sure we've completely eliminated the dependency on the CHUD framework. You shouldn't need to have it installed.

For the Windows and Linux builds, we're experimenting with using OpenMP. We aren't using it throughout Onlink much yet, but CrissCross uses it fairly extensively. This means a few things for users:
1. The Linux build now depends on libgomp.so.1, provided with GCC 4.3.3. Anything GCC 4.3 or above _should_ work.
2. The Windows build may or may not depend on the Microsoft-provided DLLs for OpenMP. If you get runtime errors before the program even launches, try installing the MSVC 2005 runtime. (x86, and/or x86_64)
3. OpenMP is going to allow us to do some pretty nice optimizations, so expect some performance improvements to be forthcoming!

Also, we've switched to an XML-based save system, which will hopefully allow us to better support upgrading old savegames to new formats. Let us hope that this is the last time we have to ask you to create new save files!

[/lurk]

Well, the new build seems solid, except for one thing: It seems I'm completely unable to delete logs of any kind. Running the x86-64 version under Win7 (x64), fresh install of course. This error (errh, unless it's some new feature I didn't hear about?) is consistently reproducible - I went trough the following process twice (including the reinstall) and the same thing happens.

Detailed reproduction steps:

• Made a fresh install of Onlink "Beta" (it did show 0.2.0-a8) using Updater (after uninstalling previous and rebooting PC).
• Started the game, went through first time registration.
• Skipped tutorial, entered Uplink Internal Services, bought Log Deleter (any version does this), then disconnected.
• Entered InterNIC, searched for "uplink", added the Uplink Test Machine to my front list.
• Setup a bounce through InterNIC ending at Uplink Test Machine, then connected.
• Logged in with the default admin password, accessed logs.
• Started Log Deleter, but when clicking on any log item with the Deleter selected, nothing happens, the logs remain there and the deleter remains attached to the cursor. No message popup either.

Oddly enough, when I went to check to see if there were any CrissCross console messages related to that issue, I found that it seems to be stuck at "Loading LANs...".

I tried to wait a bit to see if that would clear up. At some point, the autosave kicked in. I went back in-game, and saw that I had an incorrect entry for the Test Machine password (well, the password was there, but as a password-only code, the "admin" entry for the user field missing - clicking on the code produced no visible effect - thanks for working on adding that feature though!). After just waiting there a little bit more, I started getting recurrent "WARNING: .\interface\remoteinterface\useridscreen_interface.cpp ln 105: tried to use an invalid code FAIL: backtrace() function not available." messages in CrissCross.

Still fiddling around with it, I'll report back if I find more stuff.

New build! ^v^ *downloads, runs off to play*

I have had the same problem with deleting logs. I hacked my way into a bank with $1,000,000 and transferred it into my account. I had everything bypassed and was ready to delete the log. I went to the account statement and there was only one log there (my transfer), and I couldn't delete it using my level 4 log deleter. So that game got caught.

I'm having a problem as well...

On the mac build, as soon as I open onlink it quits.

We'll need a debug log.

~/Library/Application Data/Onlink

If that folder doesn't exist, tell me so I may proceed with beating Tycho with a chain.

There appears to be no "Application Data" folder in ~/Library/.

And in case you meant "Application Support," there is no Onlink folder in there.

Huh. I think I may make some money selling chains here... >_>

I did. Make the Onlink folder in there. If it magically works, it's beating time.

So, here's the deal - log deleter doesnt work, but the modifier DOES, so the problem is probably on the program, not in the logs themselves.

Also, the windows autosave lag now takes about dastard 10 seconds, instead of the previous annoing 3 seconds. This makes it hard to even play the game and try to find more wrong stuff...

Turn off the autosave... It's in the options menu.

Edit: I just had a small typo that cost me a couple minutes. I had to change a technical contact to "Christoper Jay." Note the missing h in Christopher.

I put an Onlink folder into Application Support, no debug log, and Onlink still quits.

Eh? Where are you getting that from? I'm getting 0.23 seconds of autosave for a 22MB save file as opposed to previously 0.18 for a 23MB save file. Those are, of course, the decompressed numbers - the gzipped xmls are much smaller.

And yeah, I tried that, and the log modifier does work normally. I wouldn't go so far as to assume it's definitely not a problem with the logs themselves though - maybe the records can be changed, but not erased? At any rate, I'd just let the people who actually know this code figure out what the real problem is, rather than throwing suppositions around. Betatester's 101.

I forgot to mention in the post: save times on Mac and Linux are instantaneous. If your save times are really that bad, you may want to try another OS.

The reason that saves are so much slower on Windows is that we can't just fork() and do the work in a separate process, we have to wait for the entire world to serialize into XML and then we can finally write it to disk in a separate thread.

I'm aware of which change did this. Hopefully you guys can survive without Log Deleter for a while, we'd ideally like to fix a few more bugs before we do another build. Was it stuck there, or did it just not print a time for it? If the game was responsive, then it wasn't actually still loading LANs. The x86_64 build can't do backtraces, which means we don't get useful debugging information from it. Can you try reproducing this on an x86 build, bugreporting it to Bugzilla and posting the backtrace it gives for that warning?