WinUAE 2.7.1 Betas

Printable View

6th January 2014, 13:03
Demon Cleaner

WinUAE 2.7.1 Betas

Beta 1:

Quote:

Read following instructions before posting http://eab.abime.net/showpost.php?p=861456&postcount=1

2.7.1 (or possibly higher) beta series. Lots of A500 cycle-exact improvements and chipset corner cases fixed. (Thanks to amilo3438 for collecting huge amounts of demo test cases)

http://www.winuae.net/files/b/winuae_2710b1.zip

Huge changelog, mostly technical stuff.

- Setting blitter channel mode to D only from D + any other channel(s) combination when blitter was active caused blitter emulation to repeat D write until BLTSIZE was written again. (Sinemania / Latex)
- Blitter was not fully synced with bitplane DMA. When CPU had nothing to do (waiting for memory fetch or had long internal operation executing), blitter and bitplane DMA was emulated in separate passes. Optimization which is not that good idea in cycle-exact modes. Fixes glitches in most demos with vertical "copper" bars.
- "Blitter: D-channel without nothing to do?" can happen if program changes channel mode on the fly from D to AD, do not cancel blit in this situation. (Just Another Compilation / Magnetic Fields)
- Blitter´s last idle cycle before final D write does not require free bus cycle. Exception: it needs to be free if it is extra cycle added in fill mode. (Lets Rave / Profecy, CrazyWorld / Dioxide)
- D-only blit didn´t set blitter finished state early enough. Blitter finished bit is set, in normal mode, after second to last D write is done, not when blitter pipeline is empty and final D is written. (Bob demo / D-Mob)
- Blitter line mode timing updates, blit finished bit gets set immediately when last D write has been done but copper waiting for blitter needs 4 extra cycles and they don´t need to be free cycles. I assume these 4 extra cycles are same as non-line mode 2 extra cycles + line mode forcing blitter to use only every other cycle. (Matches logic analyzer results but this is not easy to 100% confirm)
- Writing to blitter pointer register 1 cycle before same blitter channel is used for DMA fetch: written value is ignored, pointer does not change. Cycle exact mode only. (Rampage / TEK music corruption in Credits and Acknowledgements part, corrupted when landscape part started)
- Blitter not in nasty mode: idle cycles (that are usable by CPU) incorrectly counted as "CPU waits for bus".
- Blit without D-channel active in cycle-exact mode: final cycle still used D channel. (Oops..)

- When copper was stopped by writing to both COPxJMP registers while DMA was also disabled, it incorrectly was set to strobed state if yet another COPxJMP access was done with DMA still being disabled.
- Emulated quite weird copper feature, if second cycle of copjmp strobe cycle sequence would cross scanlines (positioned immediately after first strobe/refresh slot), it will disappear! (Psycho Medium 4 / Turnips)
- New undocumented feature: bitplane DMA and sprite DMA is inhibited during last line of field. (A500 logic analyzer confirmed, no difference between OCS or ECS)
- DDFSTOP > maxhpos causing full horizontal overscan DMA (ECS only) improved. (Child´s Play 1 / Silicon Ltd)
- "SWIV" bitplane feature was line based, now it is cycle-based. (No one has enabled or disabled it mid-scanline, but at least emulation is more compatible now)
- If number of bitplanes decreases mid scanline, possible remaining data (BPLCON1 non-zero) in disabled planes´ shift registers must be shifted out. (Party Report / Escape & Outlaws)
- Optimized "SPEEDUP" bitplane emulation didn´t copy last fetched data to BPLxDAT storage variable, causing random glitches in other parts of display if program changed number of planes after DDFSTRT had matched. (Plasma Explosion / Timex)
- Bitplane delay (BPLCON1) mid-scanline timing should be now 100% exact. (Psycho Medium IV / Turnips)
- Sprite update in 270b13 was wrong, should work better now, it broke Elfmania scoreboard.
- Modified interrupt delays, handle situation where (broken) program starts blitter with INTREQ already set and blit finishes before next instruction clears it. (Absolute Inebriation / VD "big double glenz" part)

- Reset color registers only after hardware reset.
- It seems ECS Denise color palette is full white after power up but full black if OCS. (Small Intro / TWW)
- Reading byte from CIA address that has valid CIA A/B select bit but bit 0 inverted (Active 8 bits of 16-bit data bus not driven by CIAs) didn´t read prefetch buffer contents in 68000 CE/prefetch mode. (Copper Bandwaggon / New Age)
- 68020 CE mode 6888x exceptions didn´t skip instruction prefetch, broke for example KS 1.3 FPU detection.
- 68020 CE last prefetch word load after branch or exception used incorrect (too fast) non-ce fetch call.

- MSVC 2013 compiled, switched to fastcall calling convention.
- Added optional gameports panel setting: don´t disable original key if it is mapped to keyboard joystick emulation. Config file only, joyportXkeyboardoverride=false (X=port number).
- Added "delay " command to custom event strings. (for example "kbr ´a´" "delay 50" "kbr ´b´" will send key ´a´, wait 1 second (in PAL mode), send key ´b´. Multiple delayed custom events can be active at the same time.
- Accept Draco Casablanca partition table identifier (CDSK instead of usual RDSK).
- 256k rom image inserted in floppy drive: mount as A1000 KICK disk, 512k rom image mounts as A3000 SuperKickstart disk (with 256k rom area empty and no bonus code included = useless for A3000 emulation)
- Resolution autoswitch was not compatible with new interlace scanline modes.
- Resolution autoswitch improved, instead of being boolean (on or off), now it can be configured for percentage of resolution used, for example if value is about 10 or larger, and game has tiny hires scoreboard and the rest is lores: lores will be selected. Value=1 emulates old behavior, use highest resolution if there is even single line of hires or shres mode.
- Some Game Ports panel remap operations didn´t work correctly if Input panel was set to "Configuration #1-#3"
- Automatically focus debugger window when debugger activates.
- Capture mouse after exiting debugger if it was captured when debugger was activated.
- Do not ask for extended ADF if program writes to disk using standard dos format and 1 or more sectors have non-zero header data, add message to load and ignore sector header data. (Thrill Kill disk writing)
- Standard partition HDF zero or -1 dostype: force DOS\0.
- <= 512M partition hardfiles had IDE-like geometry if it didn´t have OFS or FFS filesystem.
- Always mount all non-CDROM SCSI devices if uaescsi.device is enabled and SCSI mode is SPTI.
- More readable crash dump file naming, 64-build dump file support added.
- Mouse driver mode + magic mouse lost mouse buttons when WinUAE window focus was lost.
- RTG mode + Aspect ratio not default or automatic + Scale if smaller set: screen was very narrow.
- DMA debugger blitter color changed, now normal blit, fill blit and line blits are different enough.

- JSR and JMP (d8,PC,Xn) or (d8,An,Xn) addressing modes did one unneeded prefetch.
- MOVE from SR was too fast if destination was memory (68000 does memory write twice for some reason)
- SR/CCR modification instructions always do full prefetch fill.
- (d8,PC,Xn) and (d8,An,Xn) addressing modes have idle cycle before first prefetch, not after.
- CMP.X #imm,EA where EA = -(a0) or (d8,An,Xn) didn´t have extra idle cycle.
- Logic analyzer rechecked and updated exception initialization times. (Some were 2 cycles too fast)
- DIVU/DIVS cycle count algorithm had prefetch included: division operations were always 4 cycles too slow.

- Arcadia support synced with MAME 0.152 (ar_xxxx.zip rom images), missing bios roms and games added. Old rom files not supported. Note that Arcadia is NTSC, new bios 4.00 hangs in PAL mode. ROM rescan required to detect new files.
16th January 2014, 05:10
Demon Cleaner

Beta 2:

Quote:

- BPLCON1 mid scan line modifications now should match real hardware 100%.

This was very big change, old emulation was functionally close enough but very far from real hardware behavior. Old code couldn´t emulate some side-effects correctly.

- Partial/full scanline based mode ("SPEEDUP") is again used more regularly, "is it safe to use" check got partially broken during 2.7.0 betas. It is also more optimal in AGA 32/64-bit fetch modes.
- AGA border sprite in right border and 2x or 4x horizontal sprite pixels (vs bitplane pixels): sprite´s rightmost pixel(s) were missing.
- Superhires and top or bottom border sprite: graphics garbage was seen in border area, memory corruption was also possible in some situations.
- Magic mouse + mousehack mode now always stops keyboard input when mouse is outside of emulation window, even if window still has focus.
- 68020/030 CAAR register preserves also all reserved bits.
- Disable JIT if config has 68000/010 CPU and non-zero JIT cache size.
- 68020 CE broke interrupt state fetch, interrupt didn´t clear if write to INTREQ was immediately followed by RTE. (b1)
- Does not anymore require SSE2 capable CPU (b1 did)
17th January 2014, 05:17
Demon Cleaner

Beta 3:

Quote:

- Sprites near left border had strange corruption (b1)
- "toscr_nbits > 16" error in some situations (b1)
- "Modified interrupt delays" b1 change was bad, sometimes interrupt request was kept active causing spurious interrupts (which caused crashes).
- New autoresolution selection was not loaded correctly from config file.
- Blitter final D write don´t increase "nasty" count. (Which makes sense, blit is already finished, blitbusy bit was cleared 2 cycles ago)
- Removed most BPLxDAT related hacks, they become obsolete after b2 BPLCON1 update (it also changed how data gets moved to shift registers). May introduce new glitches and if it does, it needs correct emulation updates, not hacks.
20th January 2014, 13:30
Demon Cleaner

Beta 4:

Quote:

- More than one horizontal DDFSTRT-DDFSTOP window is now accurately emulated without requiring any hacks. Random garbage between display windows is also gone. (Subtle Shades / Nuance)
- Removed some more obsolete variables from display emulation.
- Saigon Megademo fix got accidentally broken in b2.
- 260b17 "strange right edge overscan display shift" optional hack removed, new delay/shift register behavior creates identical output automatically simply by reseting delay counter (hpos counter) at the end of scanline. NOTE: Some programs with large overscan can now have small graphics corruption in right border (gap in graphics), it is not a bug, it is accurate emulation. (For example Back in Bizness / 2000AD)
- Some OCS only programs may now have graphics corruption in ECS mode (if bitplane DMA crosses scanlines), this is also not a bug, real hardware would show even worse corruption. Someday this will be more accurately emulated.
- Agnus vpos counter increase delayed by 1 cycle. Fixes Hit the Road / Flash Production "Floffy2" part.
- Keyboard leds as power/floppy/etc indicators are now properly restored back to original state when keyboard input is not active, changing led config on the fly also works now.
- Exiting GUI with close button or ALT-F4 didn´t save GUI position.
13th February 2014, 13:47
Demon Cleaner

Next betas:

Quote:

Beta 7:

A4091 and A4000T NCR53C710 SCSI emulation works now! Now all built-in devices in all supported setups are emulated. I guess last interesting big missing Commodore designed part is CD32 MPEG module.
Added A4000T and A4091 to ROM scanner results window.
Added A590/A2091 and A4091 boot ROM selection to GUI, if empty, uses old automatic selection.
A590/A2091 + Fast RAM: Fast RAM is always first, A590/A2091 second, then other boards (if enabled), now matches real A590/A2091 with fast ram installed.
HID RawInput had wrong access mode in device open call, failed if other users of same device used different access modes.
68EC020 "opcode x has faulty prefetch" triggered too easily when program did stupid things. (It is only a debug feature, it will be gone someday)
AGA high fetch mode screens with ddfstop near the end of right border were completely blank.
DDFSTRT limit should go away if programmed mode or superhires (broke in some previous beta).
Reset horizontal display window if it is still active at the end of scanline (even if it is larger than last hpos value) if programmed mode or superhires. Perhaps this is wrong, perhaps not, not bothered to test yet.. (Background "leak" in right edge in some wide programmed mode screens)
Accept uaehf.device open call even if ioreq->io.Message.mn_Length is invalid or too small.
Renamed some protection dongle names in GUI.

Beta 6: (Probably fastest beta update ever.. B5 is broken in non-ce modes if chipset modes are used. Only RTG worked.)

Display emulation last horizontal position (plfright) was not correctly updated, caused all kinds of bad side-effects, including non-immediate/non-ce blitter taking seconds to finish.

Beta 5:

"Strange right edge overscan display shift" adjusted, in some situations it was visible when it shouldn´t have been. Looks also correct in NTSC mode now (Sawtooth pattern caused by alternating line lengths)
OCS Agnus requires one cycle between sprite DMA and bitplane DMA, ECS Agnus can have both cycles back to back. This means OCS Agnus will only fetch first sprite word (practically making sprite useless) where ECS Agnus can fetch both.
Rarely during heavy interrupt activity in cycle-exact mode some delayed interrupt register set or clear operation may have been lost causing possible spurious interrupt.
lea (d8,An,Xn) prefetch is last operation, not idle cycle.
68020 CE mode CAS instruction prefetch was broken.
When UAE boot ROM is enabled, add fake autoconfig board (configdev) that has same board address and size as boot ROM. Allows MMU utilities (mmu libraries, enforcer etc..) to mark boot ROM as valid address space without extra command line parameters.
Partially rewritten horizontal Agnus bitplane state machine. Can (as usual) break something..
When BPLCON1 matches, only copy enabled planes to shift registers. (Back In 90 / Samplers)
Power led didn´t turn off after soft reset.
Added sprite-only linetoscr routines, bordersprites and HAM do not mix well. (Satyanarayanan / Scoopex)
AGA and 64-bit fetch modes: last 16 or so pixels were not flushed to display, depending on BPLCON1 value. (Satyanarayanan / Scoopex)
New (still WIP) modern native library interface merged from FS-UAE.
Added LSI 53C895A SCSI chip emulation from QEMU. Lots of changes to make it 53C710 compatible (53C720 and later have multiple registers changes but SCRIPTS code is fortunately backwards compatible) A4091 and A4000T NCR driver now "work", chip init works, returns timeout error (no device connected) when it tries to detect SCSI devices. Do not connect any drives, it does not work yet.
pdk compressed (or dsq) files always unpacked to Amiga DD floppy size.
Keyboard leds lost sync in some situations.
25th February 2014, 12:50
Demon Cleaner

Beta 8:

Quote:

- 68000 CE/prefetch mode reads of non-existing memory should return correct data in all situations.
- Reading of write-only register will now write correct data to register. Data that CPU gets back still isn´t 100% correct (which can be totally different than what was written to register).
- ROM image didn´t load correctly if ROM image was in multiple parts (odd/even) and merged parent ROM was not found.
- NCR SCSI: SCSI commands that failed returned wrong status.
- uaehf.device SCSI Verify (10) implemented. (Does not verify anything)
- A590/A2091 and A4091 options moved to Expansions panel, they are expansions, not built-in features and not part of internal hardware config (advanced chipset). CDTV SCSI is also technicallybuilt-in because driver and DMAC are located on mainboard, CDTV adapter only contains WD SCSI chip.
- A590/A2091 linear ROM mapping was incorrect (Any CPU mode without "more compatible" fetched bad opcode data). Note that A590/A2091 can´t be JIT compatible.
- Added "ROM disabled" option to A590/A2091 ROM select menu. (ROM disable jumper is for 1.2 compatibility)
- It was possible to set and get unused DMACON bits 11 and 12.
- A4091 was not compatible with Z3 Fast.
- Hardcoded A4091 autoconfig data was replaced with data from ROM image.
- Border blank didn´t override AGA border sprites (broke in some previous beta).
- 68020/68030 SR M-bit should be cleared (if set) before executing interrupt routine.
- 68030 MMU FPU FMOVEM instruction continuation state recovery after bus fault was incomplete.
- Generate bus error when accessing invalid memory if A3000/A4000 Fat Gary based system, MMU emulation active and Gary TOENB register bit is set.
- Joystick/joypad can be used to control light pen cursor. (Was mouse only previously)
- Some kind of physical mouse as digital/analog joystick/joypad support. Not that good but perhaps better than nothing.
- Removed useless address space gaps between Z3 boards.
- CD image CDA player may have gotten confused in some situations if new track was requested while other track was still playing.
- 6888x Packed-Decimal Real with Dynamic k-Factor data type was detected as invalid instruction. Allowed now but k-Factor is ignored (Packed-Decimal Real with Static k-Factor was also ignored) but it isn´t that
important, practically no one uses packed decimal data types.
- Some CPU and FPU emulation changes to make it more usable with other emulators (I am working with author of Next emulator Previous to get 68030 based Next computer booting)

Beta 9:

Quote:

- NCR emulation didn´t work with SCSI IDs 4-6.
- NCR SCRIPTS MOVE command phase bit test mismatch returned wrong interrupt status. (NetBSD 1.x)
- Added partial hack to NCR emulation that "fixes" NetBSD harmless "Phase mismatch: REQ not asserted!" warnings.
- IDE emulation: use LBA48 calculations only if command is new LBA48 command (read/write ext). This could have caused unexplained disk error messages if device driver used both lba28 and lba48 commands and HDF was lba48 enabled (>=128G in size)
- maxmem command line parameter can now also be used to increase natmem limit (if host OS allows it)
- b5 interrupt update had another problem, CE modes lost interrupts in some situations when using UAE HD controller.
- Partition hardfile dostype was set to mostly random value with some CPU configs (b1).
- Disabled "smart centering" (tries to keep old centering parameters if it is "good enough"), it does not seem to work correctly anymore (probably worked only accidentally) and leaves normal WB screen modes non-centered. Will be debugged more if something else breaks, as usual.
5th March 2014, 20:23
Demon Cleaner

Beta 10:

Quote:

- Programmed display mode positioning bug fixed, should fix remaining problems with clipped left or right edges.
- Switching between interlace PAL and interlace NTSC modes calculated wrong vertical line count.
- Some CPU emulation optimization, only modify linear memory address in getpc/setpc/incpc functions if mode is not prefetch/ce/mmu.
- Added slirp inbound port configuration, config file slirp_redir=:::, multiple entries supported, internal ip part is optional.
- Added simple slirp inbound port configuration, config file slirp_ports=port1,port2,.. (TCP ports only, internal address is default 10.0.2.15, host port equals internal port)
- Added quick slirp inbound port selection to GUI, opens TCP ports 21, 22, 23 and 80. (Same as slirp_ports=21,22,23,80)
- Added missing slirp cleanup code from QEMU.
- 6888x Packed-Decimal Real negative k-factor values fully emulated (Previous confirmed, required by NeXTSTEP OS), positive non-zero values not emulated yet which needs Fortran "E" scientific notation format string that C does not directly support.
- First and last drawn scanlines were always detected incorrectly since few major versions ago, breaking vertical centering.
- Input device autoswitch by using analog directions now only work with X or Y axis, other directions may work differently and can cause accidental switches.
- Force recalculation of horiz and vert centering (if enabled) when screen mode is detected (guessed) as changed.
- Toggle input events (warp mode, mute etc..) now also support setting to on or off state without toggling. Input panel only, click on input event name to change mode
- Added audio/video recording input event, supports also new toggle/on/off modes. If recording is started using input event (not from GUI), number is appended to filename and it is increased each time recording is started.

SLiRP inbound ports note: all ports will be opened when network hardware is initialized, there is no way to know when something actually starts listening for some port. This also means you won´t get exactly same behavior as "real" low level network emulation, mainly you won´t ever get connection refused if port is open but not actually listening on the Amiga side, connection opens and then closes immediately.
25th March 2014, 07:01
Demon Cleaner

Beta 11:

Quote:

- Borderblank + bordersprite caused border glitches in some situations.
- CDTV front panel buttons´ default state after reset should be enabled, not disabled.
- Added new double options to non-lace modes. Not really useful, maybe be moved somewhere else later..
- 68040 MOVE16 write access error stack frame is now correct.
- Editing non-existing new style real harddrive config entry opened hardfile properties dialog.
- Separate filter options for native and RTG modes. Currently only following settings are RTG compatible: Direct3D shader filters (no software filter support and never will be), masks, overlays, scanlines.
- New automatic resolution option was unreliable in some situations.
- Do not detect modes with VTOTAL=350 or less as scandoubled to fix Super72 halved height.
- tablet.library emulation, totally untested, do not enable.
- After long field to short field change last line still showed old background color(s) if line had any background color changes before field mode was changed.
- Master floppy write protection config option was not loaded from config. (wrong config entry name..)
- Disk emulation multirevolution support was unstable, disk revolution data stream change was not always bit perfect.
- Added Super Card Pro (scp) image file support. (Don´t ask me if images don´t work, I didn´t implement it originally)
- Added support for raw dumps from Softpres Capture Tool and KryoFlux using new capslib v0.5.

Yes, above two changes in same beta is only a strange coincidence. Note that both are raw dumps which can´t be confirmed easily, don´t bother me if there are compatibility problems.

FPU improvements (Mostly low level stuff that only makes difference if unimplemented checkbox is not ticked, much more important for Previous emulator, NextStep OS requires FPU and uses it heavily):

- 6888x any FPU instruction (except FSAVE/FRESTORE) causes null state to idle state transition but 68040+ conditional FPU instructions won´t affect null state.
- 68040 unimplemented FPU instruction emulation was broken, FSAVE exception frame was not emulated, 68040.library emulation code needs it for proper operation.
- 68040/060 FPU FMOVECR is now unimplemented instruction.
- 68040/060 unimplemented packed-decimal data type exception is now emulated.
- 68060 FMOVEM.X with dynamic list now causes unimplemented effective address exception.
- 68060 F.X #imm,FPn and F.P #imm,FPn now causes unimplemented effective address exception.
- 68060 FMOVEM.L #imm, now causes unimplemented effective address exception.
- FMOVEM.X with dynamic list and -(An) and (An)+ effective address overrides static/dynamic post/pred state. (for example FMOVEM.X Dn,-(An) uses pre-decrement even if mode bits are set to dynamic post-increment)
- All FPU FPSR register (overflow, div by zero etc..) status bits emulated. (May not be fully reliable)
- 6888x Packed-Decimal Real k-factor support fully implemented. (May still have rounding bugs)

This should complete emulation all FPU features except arithmetic exceptions (that no one seems to use)
30th March 2014, 06:17
Demon Cleaner

Beta 12 & 13:

Quote:

Beta 13

Canceling video or audio codec selection dialog will disable audio or video recording. This has been broken for long time, it was impossible to record video only or audio only (except in wav mode).
64-bit build now also supports wintab tablet API.
Fixed some slirp Windows specific 64-bit compilation issues.
Fixed JIT + more compatible enabled emulated Amiga crash. (b9)
If "always on top" option is enabled, GUI is also opened with "always on top" flag.
gfx_linemode=none (no doubling) was loaded as doubled mode.
Check and limit 2x software filter height value if size is larger than allocated buffer size, preventing display buffer overflow crash.
Debugger now disassembles most FPU instructions correctly but some of them may look strange..
Disassembler MULL and DIVL fully decoded.
Support "INPUTDIMS" D3D shader variable. Used by some shaders.
JIT FPU rare bug workaround, FPU constant optimization introduced in 1200b3 disabled. (Code looks correct, it works correctly except in some very rare situations, real cause may be some unknown JIT side-effect)

Beta 12

Config file writing created corrupted (was mostly harmless) config file lines, most common being uaehf line which fortunately isn´t important config line entry. (b11)
BCHG/BCLR/BSET allowed PC-relative addressing modes, broken since the beginning. (from Hatari)
If program read disk by polling DSKBYTR (not using DMA), index emulation was unstable.
Switching bitplane DMA on few cycles (not yet confirmed, I assume range that makes sense is 1 to 3) cycles before DDFSTOP will also trigger ECS/AGA "passed plfstop" state change. (Death Trap)
Full tablet emulation boot division by zero guru fix.
Tablet.library emulation now works but at least dpaint5 does not seem to do anything with pressure data. It reads pressure tag contents but nothing seems to happen.
15th April 2014, 17:47
Demon Cleaner

Beta 14:

Quote:

- FSINCOS second register supported in disassembler.
- FMOVEM and FMOVECR disassembler support. Disassembler should now support all FPU instructions.
- Added missing special case 68020+ instruction disassembler support: bit fields, MOVES.
- Disassembler output improved, most unsized instructions don´t have .L size extension anymore.
- Action Replay 2/3 was incompatible with b8 cpu emulation changes.
- CPU mode (JIT/more compatible/cycle-exact) on the fly switching was unstable since b8.
- Mirror Expansion panel RTG scale options in Filter panel. (More options will be added in future)
- Removed some forced inlining that only increased size of executable and wasted CPU cache space.
- Added support for early 68040 revision that has slightly different FPU stack frames and version id. (Config file fpu_revision=0x40). Not much use in Amiga emulation but first 68040 based Next systems came with first revision 68040´s and NextStep 2.0 unimplemented emulation code only supported first revision unimplemented stack frame type.
- FMOVEM predec/postinc mode bit emulation fixed again, now it really works correctly.
- Implemented (partial) 68040 FPU BUSY stack frame.
- Pre filters in RTG mode loaded incorrectly from config file.
- Implemented FPU exception state statefile support.
- b13 JIT FPU bug workaround replaced with fixed optimization (Peter Keunecke)
- Attempt to automatically select correct resolution (hires/shres) when programmed mode is active. Always active if resolution setting is hires or larger and doubling is enabled. No more programmed modes that are 2x wide or half width without manually changing resolution setting, note that currently "wrong" mode is visible at least 1 frame before mode changes.
- Arcadia game Aaargh supported.

Beta 15: (Should be feature complete now. Perhaps)

Quote:

- Finally bumped version to 2.8.0. This isn´t small 2.7.0 update (Should have been done since b6 or so..)
- FPU register negative zero transformed to positive zero when it was written to or read from memory in extended FP format (FMOVE.X FPx, or FMOVE.X ,FPx or FMOVEM). No one should care as usual.
- When loading FPU register from memory in extended format, normalize any "unnormal" zero. (non-zero exponent but zero mantissa part). Motorola FPUs do this automatically, x86 FPU does not seem to like them. (Another mostly Next specific update)
- Programmed mode autoresolution should now work with strange modes like highgfx.
- Only allow CPU mode changes between instructions, previously change sometimes happened mid-instruction (during instruction internal memory read/write) which wasn´t safe.
- Small tweak to recent sprite update (Hybris)
25th April 2014, 07:38
Demon Cleaner

Beta 16:

Quote:

- Borderblank blanked sprites incorrectly in some situations (b11, Roots AGA)
- SPRxDATx/SPRxPOS copper write/Denise internal sprite timing should now fully match real hardware.
- Ancient (and recently non-working) Superfrog flashing intro bee hack removed, properly emulated now.
- gfx_filter_keep_autoscale_aspect and gfx_filter_keep_aspect config entry change was not detected when using using uae-configuration.
- Last used AVI codec was not restored from registry if ICGetState() returned zero size data block.

Beta 17:

Quote:

- Debugger breakpoints work again. (b14)
- Full CPUSHA/CPUSHL/CPUSHP disassembly implemented.
- Integer scale mode compatibility improved with programmed display modes.
- Integer scale mode changes Filter panel zoom slider range to -99 to 99 and is used to adjust integer scaling calculation. For example -10 means if next largest scaling factor is selected even if display height is 10% too big to fit normally. (This was implemented already but it didn´t work properly)
- Fixed fill_line() top/bottom background line drawing buffer overflow in some programmed modes.
- Some programmed modes had single short background colored line in left horizontal blanking area.
- Command line (or icon doubleclick) loaded config didn´t reset previously loaded config (default.uae, if it existed)
- Emulated more Fat Gary address regions that are unmapped (CIA is same as Gayle, upper 32k in clock space and Gary register space are unmapped, 0xdf0000-0xdf7fff is also unmapped)
- Emulate Fat Gary invalid address space access timeout if Gary timeout bit is set, first 50 timeouts are always logged. To prevent last moment stupid bugs or bug reports, delay is not yet