flashjazzcat Posted February 19, 2015 Author Share Posted February 19, 2015 Added to the to-do list for today. Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 19, 2015 Author Share Posted February 19, 2015 (edited) I'm not yet sure, if it is a software bug or my actual Atari hardware. I'm using the Atari here and then, but it never felt "unstable" to me.It's running nice for most of the stuff I'm doing like playing games and watching demos. Interesting - thanks for posting these. They're very "early" crashes: normally MrFish and I had to open and close things for quite some time in emulation in order to trigger problems in the pre-release builds, which was good since I was then able to target memory and run-queue deallocation as the probable source of issues. Not here, however. It's not insufficient RAM, since the issues in the video are not reproducible on an emulated 130XE. Problems seem to commence on your set-up as soon as extended RAM is allocated (once the main bank has filled up). All ROM bankswitching is performed from code in RAM, so I can't envisage any timing issues with the cartridge being the culprit. PORTB is worked very hard once multiple applications are running, so although I have no clue whether this is a software or hardware problem, for sure any RAM problems will jump out when running the GOS. In both cases, what we seem to be seeing is corruption in the window control records, which cause the UI manager to crap out after drawing the tab control in the last opened window. So: something went wrong during or after the profiler was loaded and relocated in extended RAM. Edited February 19, 2015 by flashjazzcat Quote Link to comment Share on other sites More sharing options...
+Ripdubski Posted February 19, 2015 Share Posted February 19, 2015 SIDE2 will ultimately be your best bet for this, I'd say. Built-in HDD and RTC. Minus RTC, MyIDE II ? Quote Link to comment Share on other sites More sharing options...
+MrFish Posted February 19, 2015 Share Posted February 19, 2015 Interesting - thanks for posting these. They're very "early" crashes: normally MrFish and I had to open and close things for quite some time in emulation in order to trigger problems in the pre-release builds, which was good since I was then able to target memory and run-queue deallocation as the probable source of issues. Not here, however. It's not insufficient RAM, since the issues in the video are not reproducible on an emulated 130XE. Problems seem to commence on your set-up as soon as extended RAM is allocated (once the main bank has filled up). All ROM bankswitching is performed from code in RAM, so I can't envisage any timing issues with the cartridge being the culprit. PORTB is worked very hard once multiple applications are running, so although I have no clue whether this is a software or hardware problem, for sure any RAM problems will jump out when running the GOS. In both cases, what we seem to be seeing is corruption in the window control records, which cause the UI manager to crap out after drawing the tab control in the last opened window. So: something went wrong during or after the profiler was loaded and relocated in extended RAM. Unfortunately I haven't been able to try this out on real hardware yet, as I sold off my only MaxFlash cart when I let go of my Pac-Man Arcade copy. I was using it to flash other MaxFlash images before. I can get another one, but once you start making SIDE images I can start using them without having to purchase anything. I thought the same thing about the extended RAM usage during his crash. I wonder if his extended RAM is alright? Any other tests to see that the extended RAM in the machine is ok, a game that uses it or a RAM testing app? Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 19, 2015 Author Share Posted February 19, 2015 Minus RTC, MyIDE II ? Certainly, and I own one, which helps. Unfortunately I haven't been able to try this out on real hardware yet, as I sold off my only MaxFlash cart when I let go of my Pac-Man Arcade copy. I was using it to flash other MaxFlash images before. I can get another one, but once you start making SIDE images I can start using them without having to purchase anything. I thought the same thing about the extended RAM usage during his crash. I wonder if his extended RAM is alright? Any other tests to see that the extended RAM in the machine is ok, a game that uses it or a RAM testing app? I'll have to get into those other images tonight since I'm busy getting a machine ready to post tomorrow, but it's the first thing I'll do when I get a chance. I suppose it wouldn't hurt to run XRAM or similar on the machine. As I was saying in my email, extended memory is typically formatted as a RAMdisk and then forgotten about, and there aren't a vast number of applications which run actual code out of extended banks. Quote Link to comment Share on other sites More sharing options...
+MrFish Posted February 19, 2015 Share Posted February 19, 2015 I'll have to get into those other images tonight since I'm busy getting a machine ready to post tomorrow, but it's the first thing I'll do when I get a chance. No rush, the demos already out. I suppose it wouldn't hurt to run XRAM or similar on the machine. As I was saying in my email, extended memory is typically formatted as a RAMdisk and then forgotten about, and there aren't a vast number of applications which run actual code out of extended banks. Yeah, I run XRAM and a couple of these games that use extended RAM: Games - Extended RAM.zip Quote Link to comment Share on other sites More sharing options...
twh/f2 Posted February 19, 2015 Share Posted February 19, 2015 Are you guys referring to XRAM 0.21.(XI/Satantronic) eXtended RAM test ? ( http://satantronic.atari.sk/xXRAM.htm ) I ran XRAM-test several times. All I can say is, that it's not reporting any errors. It simply says: "Tests done!" on a green background.. Looks ok to me... Then I tried some RAM-hungry games too: * Bombjack (320K) (fine) * Command+ (320K) (fine) * Yie-ar-kung-fu (320K) (fine) Are there any other "mean" bank switching tests around? Which I could give a try? Maybe that's the GUI's first serious application - a REALLY demanding RAM-stability testing tool -twh 1 Quote Link to comment Share on other sites More sharing options...
w1k Posted February 19, 2015 Share Posted February 19, 2015 XRAM 0.21 - http://atari.turiecfoto.sk/soft/index.html Quote Link to comment Share on other sites More sharing options...
+MrFish Posted February 19, 2015 Share Posted February 19, 2015 (edited) Are you guys referring to XRAM 0.21.(XI/Satantronic) eXtended RAM test ? ( http://satantronic.atari.sk/xXRAM.htm ) I ran XRAM-test several times. All I can say is, that it's not reporting any errors. It simply says: "Tests done!" on a green background.. Looks ok to me... Then I tried some RAM-hungry games too: * Bombjack (320K) (fine) * Command+ (320K) (fine) * Yie-ar-kung-fu (320K) (fine) Are there any other "mean" bank switching tests around? Which I could give a try? Maybe that's the GUI's first serious application - a REALLY demanding RAM-stability testing tool Yes, that's the one, XRAM 0.21. That, testing a few games, and the built in self-test are all I usually rely on. Good to see it running on some real hardware -- even if it's crashing. I didn't know FJC hadn't run it on real hardware until today. He was probably too burned out coding to test much of anything. It's just a demo anyway, but I'm sure he won't let any serious bugs linger for long. Good idea about a RAM testing app. Thanks for the help! Edited February 19, 2015 by MrFish Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 19, 2015 Author Share Posted February 19, 2015 (edited) Good to see it running on some real hardware -- even if it's crashing. I didn't know FJC hadn't run it on real hardware until today. He was probably too burned out coding to test much of anything. It's just a demo anyway, but I'm sure he won't let any serious bugs linger for long. What's puzzling is that the GOS crashes on twh/f2's machine pretty much immediately and in a predictable way. I hadn't tested the latest build on real hardware until Wednesday (following the crash report), but I did not experience any of the issues reported (nor indeed any issues at all during a brief test). Although this is a demo, it should not crash after launching only a couple of applications. The fact it does for one user is a cause for great concern, despite the fact the OS is far from complete. The tricky part is duplicating the issue. If I'd launched the thing on real hardware and it had crashed after launching a couple of apps, the situation would be very different. Good idea about a RAM testing app. Way back in time I remember Candle enlisted me to write a RAM testing application which employed various deep-testing algorithms (I implemented it based on pseudo code found somewhere on the Net). Unfortunately I think it's long gone now. EDIT: Here is the GOS running on the 1200XL on Wednesday evening: Edited February 19, 2015 by flashjazzcat Quote Link to comment Share on other sites More sharing options...
+JAC! Posted February 19, 2015 Share Posted February 19, 2015 Currently BRK is unused and unhandled from what I've seen in the code. My recommendation (seriously) would to installing a kind of "Blue Screen" to the BRK handler. Programs that crash like shown in the video, typically run into a BRK very soon, allowing you to back trace the PC. 2 Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 19, 2015 Author Share Posted February 19, 2015 That's a good idea. Quote Link to comment Share on other sites More sharing options...
gozar Posted February 20, 2015 Share Posted February 20, 2015 Tried it out on: 1200XL with U1M Stock 130XE 800XL with Rambo 256K No problems at all, I was easily able to hit the process limit and let it run. The 1200XL froze up once, but I couldn't reproduce it. One minor gui refinement, I miss the menu item flashing three times upon clicking. I usually wasn't sure if I actually selected the menu item, but once I realized there were only two items that did anything, that helped (the other menu items are just a big tease :-). A little psychedelic if you're not on svideo :-) (I know there is nothing you can do about artifacting, I just thought it was funny.) 1 Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 20, 2015 Author Share Posted February 20, 2015 (edited) Thanks. I think there's still a bug in the wake up queue, then. Menu items which aren't active will be greyed out in the final product, so things should be clear without flashing menu items. I mentioned in the blurb that only two menu items were operative for now. I do not miss artifacting. Edited February 20, 2015 by flashjazzcat Quote Link to comment Share on other sites More sharing options...
Pokey800 Posted February 20, 2015 Share Posted February 20, 2015 hello, this is great gui, really fast. i played with it in altirra but sometimes it look like it is frozen. i have the video here 1 Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 20, 2015 Author Share Posted February 20, 2015 Excellent: I've been trying to duplicate that issue for some time and at least it's reproducible in emulation. Looks like everything was temporarily stuck in sleep. Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 20, 2015 Author Share Posted February 20, 2015 Just found a bug in the masking which crops up when a selected list item is behind other windows: Still trying to trigger the non-responsiveness bug. Quote Link to comment Share on other sites More sharing options...
TheNameOfTheGame Posted February 21, 2015 Share Posted February 21, 2015 Just found a bug in the masking which crops up when a selected list item is behind other windows: highlight.png Still trying to trigger the non-responsiveness bug. Oh nasty bug! Glad to see so many testers reporting in. When the Side2 image is out I can run it also on hardware. Keep up the good work! Quote Link to comment Share on other sites More sharing options...
billkendrick Posted February 24, 2015 Share Posted February 24, 2015 Here's a ROM for a 1Mbit MaxFlash cart: A8 GOS - AtariMax 1Mbit, ST Mouse.zip ST mouse required. I just seem to be pottering around looking for bugs at the moment, so I figured why not put the thing out there, warts and all. I'll make more builds as an when: just want to start the ball rolling here. You can use the MaxFlash cart studio to make a flasher ATR if you have a real cart, or just mount the ROM in Altirra (choose MaxFlash 128K / 1Mbit). Tried it and it is AWESOME. And I will be demoing it at Atari Party in May. (I'll try to keep an eye out for any new revs between now & then, but in the meantime, this is incredible.) I don't remember how fast or slow older Macs were that I used (really early one circa 1986, and Mac IIs in the early 1990s). But I swear, this is faster than they were & redrawing 1 Quote Link to comment Share on other sites More sharing options...
Muzz73 Posted February 24, 2015 Share Posted February 24, 2015 Nice! I'll be looking for it! I'd love to see it in action up close. Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 25, 2015 Author Share Posted February 25, 2015 Been trying to make a SIDE2 build and have been derailed by the fact the banking register at $D5E1 appears to become inactive when SDX is disabled. I realized Ultimate and Incognito suffer from this, but I thought the banking register remained active on SIDE2 even when the switch was in the "up" (loader) position. Altirra's emulation suggests this is NOT the case, however, and I've tested a real cart and it crashed so I'm assuming the emulation is faithful. Puzzling, since I thought the XEX loader was a 16KB bank-switched ROM. I'll try to get in touch with Candle about it. Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 26, 2015 Author Share Posted February 26, 2015 I'd completely forgotten about the secondary banking register at $D5E4 (a D'Oh! moment - thanks Phaeron), so the build occupying the upper half of the SIDE2 cart is now working. UFlash is also getting some updates to flash all supported devices, so expect the following in the next few days: SIDE SIDE2 AtariMax 8Mbit Sic! MyIDE/Flash MyIDE II 4 Quote Link to comment Share on other sites More sharing options...
TheNameOfTheGame Posted February 26, 2015 Share Posted February 26, 2015 1 Quote Link to comment Share on other sites More sharing options...
Stefan Both Posted February 26, 2015 Share Posted February 26, 2015 ... , so expect the following in the next few days SIDE SIDE2 AtariMax 8Mbit Sic! MyIDE/Flash MyIDE II yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes, yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes, yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes,yes, yes Stefan Quote Link to comment Share on other sites More sharing options...
flashjazzcat Posted February 26, 2015 Author Share Posted February 26, 2015 (edited) Here are the promised ROMs: GOS ROMs.zip GOSMF1MB: MaxFlash 1Mbit GOSMF7MB: MaxFlash 8Mbit GOSMFIDE: MaxFlashIDE (1Mbit) GOSMYID2: MYIDE II (512KB) GOSSIDE: SIDE/SIDE2 external cart (replaces XEX loader, 256KB) GOSSC512: Sic! Cart (256KB, occupies every other 8K bank at $A000) The MaxFlash ROMs have ATR flashers I prepared with MaxFlash Studio. The SIDE ROM can be flashed to SIDE1 or SIDE2's external cart (replacing the SIDE loader) using the attached test version of UFlash (I simply lacked the time to prepare ATRs using Trub's flasher this evening, but I will do so at the weekend). The Sic! Cart version can also be flashed using the accompanying build of UFlash, as can all the MaxFlash carts for that matter, should you fancy testing UFlash instead of using the ATRs provided. uflash12Test3.zip Beware: while the attached build of UFlash has been tested well with SIDE2 on real hardware, Sic! and SIDE1 have been tested only in Altirra (no problems). All the GOS ROMs have likewise been tested in Altirra (no bug fixes have been undertaken, so this is the same build I put out over a week ago) and appear to work. The SIDE build will wipe out the XEX loader on the cart, and the attached version of UFlash requires a 256KB binary for the "external cart" slot (i.e. the loader slot), so if you want to reflash the loader, dump it to disk first so you have a 256KB padded version of the external cart's contents. UFlash allows selective flashing of the odd/even banks of Sic!, and for the GUI you should select the $A000-$BFFF (SDX) slot. If you have problems with the flasher, please report them here. This is a rough-and-ready build designed to provide proper support for the external SIDE cart, and I don't know what's broken - if anything - in other areas (Ultimate, Incognito, etc). I should add that to flash the SIDE(2) builds, you should put the SIDE's switch into the SDX position, and if you have Ultimate 1MB, turn that upgrade's SDX and SIDE Hardware OFF in the BIOS. Boot SDX off the cart, and you can then run UFlash, which will auto-detect SIDE2 (it won't auto-detect SIDE1: you'll have to go into Device->Open). Once the cart's flashed, put the switch into the "loader" (not the SDX) position (keeping SDX and SIDE Hardware disabled in Ultimate if you have one), and power on the machine. You should go straight to the GOS. To get SDX back, just put the switch into the SDX position again. Edited February 26, 2015 by flashjazzcat 6 Quote Link to comment Share on other sites More sharing options...
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.