AGEMAME HQ

I’m not quite sure how two entirely different compile scripts were able to fail together, but fail they did, so I’ve replaced the patch on the downloads section. If anyone still needs to revert with patch -p0 -E -R , the old file is here.

Huge code changes here, please report any bugs.

MPU3 is deliberately disabled as I need to do more with it to make it work.

There are some issues with the AWPVid system, related to the lamp init system, there are temporary fixes in place.

EDIT: Pending the result of the POKEY problems with mainline MAME, and a few little glitches I’ve discovered in AGEMAME, the ‘m’ patch for the limited source build has been withdrawn. Those who have applied it should reverse it by using ‘patch -p0 -E -R’, until efforts can be made to fix the diff.

I’ve copied this from the AGEMAME forum, but it’s worth making clear.

Over the last couple of years, the way AGEMAME has been released has been just about good enough to cope with the various changes MAMEDev have thrown at it in terms of packaging and coding style. With reports that we’re going further down the multi-platform route, however, this is no longer the case.

A few things are definitely going to be different starting in this development cycle:

1.The version number will become that of MAME’s own, i.e. there will be no version 0.45, we’ll go straight to 0.117.

2.Maintenance of so-called ‘forked’ drivers (i.e. drivers that have been taken from MAME and stripped of non-AGEMAME games will cease, the MAME originals will be used and AGEMAME’s own filter code will remove the unused titles)

However, here’s the one thing I want to do, that is going to be controversial.

I want to remove the concept of releasing AGEMAME in a ZIP with limited MAME source code - it’s a hangover from the days when I rewrote every necessary file, and by removing all forked drivers and code, there won’t be a need to maintain AGEMAME in a separate source tree any more.

What would happen instead is up to you, I can either release the source as just the AGEMAME source overlay, or we can go down the MESS route of having the whole MAME source included as well, like it currently is on the source browser (which would be a much bigger file size).

Personally, I can see benefits to either option, which is why I’d like to hear your opinions on the forum (there’s a sticky post there for you all to look at and respond to).

I’ve been speaking to Atari Ace, and he’s willing to make my rather messy 6840 driver into something a darn sight more palatable. Meanwhile, I’ve been starting to update the MPU4 documentation I keep on the site, with a view to describing the whole setup, and moving on to other, similar hardware made by other manufacturers (See the technical information section, it’s too big for a post here).

I managed to sort the problem out, and it makes me wonder how much of a hack the previous version was anyway.

The source cleanups certainly seem to make everything cleaner to work through, a good job as I really need to update our tech docs soon.

The recent tidying of various drivers has caused problems for me, particularly as the dot matrix routine in one driver in AGEMAME relies on the Scorpion2_SetSwitchState function which has now been made static. This is ultimately a problem of my own making, not MAME’s, but a solution will be reached eventually. Regrettably, despite these games not running 100%, something like this is considered a release blocker, as it inhibits compilation.

Speaking of compilation, I’ve taken the opportunities that u2’s new makefile offers to push more of the AGEMAME specific stuff into the ancilliary code, just to keep things tidier. Basically, makefile.age now only contains the code necessary to show the AGEMAME details on windows platforms, instead of the MAME details - you could now very easily use the regular makefile, and ‘TARGET=age’ - which now means that SDL support is 100% rewrite free (just set up SDLMAME as normal, with it’s makefile, and use the AGEMAME source only download on top, making sure that your TARGET is age).

All this, possibly tonight, still, not as exciting as CPS3, is it?

The new ones are a little easier for me to get on withith, which is giving me some ideas as to how to make SAGEMAME totally maintenance free. For now, we still need a special version of the overlay.

I’ve given up on a cut down driver for Neo-Geo, since there’s little scope for stripping anything out while maintaining the new higher accuracy.

Regression testing is on hold, due to the fact that an awful lot of drivers aren’t playing nicely with the reporting tool at present. This means I’ll probably wait until 0.45 to do anything with this, and just make a lot of changes this cycle in a bid to fix it.

Speaking of which, if Atari Ace is reading this, contact me, I want to discuss some 6840 test results that could go into the fixed driver to improve it a little.

I’m aware there’s a new patch out, but I’m going to work on that over the weekend. Right now, I want to talk about what’s happening with the AGEMAME website.

Firstly, comments have been removed again, since no-one used them, and they were just causing spam problems. This has led to a small restructuring of the site templates, which has been tested.

Secondly, we now have an Emergency Alert style system which can post messages on top of the main page and keep them there for a specified period of time, this may be handy if there are problems I want to point out, or maintenance scheduled.

Thirdly, I’ve moved as much of the static content out of the databases and into conventional PHP files as possible. While the changes will appear minimal, you should see faster load times and this site will put a lot less strain on the SQL server (all of ZTNet’s SQL is run on the same machine, so it pays to keep usage low). Some of the content has also been rewritten during the process, but that’s ongoing.

Finally, Bug Reports are no longer carried here, as again, nobody used them. I hope to introduce automated regression testing to the AGEMAME build process in the near future using the MAME tools, and report those findings here instead.

If you haven’t downloaded in a while, I suggest you do so, as the codebase has improved dramatically.

This mostly helps MPU4, as I’m focussing on that right now.

EDIT: You may have noticed we’ve experimented with including the full MAME source in the source browser, this admitedly duplicates some of the MAWS data, but may prove handy to some who want to make direct comparisons within the AGEMAME framework. Personally, the jury’s out as to whether I keep it this way, as it takes up a lot more space than the old implementation.