MAME Changes that relate to FME

[Altharic]

Anyone able to try on a pc with a touchscreen does it pick up the touches if the windows one doesn't the android one won't?

[johnparker007]

23 minutes ago, Altharic said:

Anyone able to try on a pc with a touchscreen does it pick up the touches if the windows one doesn't the android one won't?

Probably a good shout - though could be how the mame4droid java wrapper is passing through the mouse data, though I'm not really up for implementing this part to be fair if not!  I'm certain it could be hooked up though with a bit of coding on the mame4droid side of things  

I recorded a terrible quality video showing indy running on my Fold4 , though having to workaround to control some of the buttons - WARNING heavy breathing in this video!  To be fair I'd just eaten and I was doubled over one phone holding another phone, sat in bed, it was all really awkward, I'm not actually that out of shape  

 

[Altharic]

You could post it as an issue on the git and maybe he would add it?

[jabbathehut]

Looks good unfortunaltly i am incompentant in setting it up properly but it does have touch controls that work.

Easier to use retroarch tbh.

 

 

 

[Altharic]

Retroarch is Satan's emulator did the buttons work in game for you?

[johnparker007]

13 hours ago, Altharic said:

You could post it as an issue on the git and maybe he would add it?

Good shout, I have opened an issue, as I'm not up to coding owt at the mo, he might be up for fixing   If he does, I will then look into a test of an old mfme2mame dx... though I did notice when setting the internal render resolution to native (from 640x480), the speed was very slow... so this won't ever be a satisfactory android DX FME experience, while the rendering is so very slow.  This was on my Fold4, which I treated myself to, so it's a pretty modern phone... Oasis will probably still end up being the way to play high res DX layouts on mobile tbh... interesting to tinker with this thing though  

[johnparker007]

Just an update to this, he said he didn't realise how powerful the artwork system was (i.e: playable slot machine layouts), so he's going to take a look into it when he gets some free time   Issue is here:
https://github.com/seleuco/MAME4droid-2024/issues/1

 

[Altharic]

I actually had a play with Retroarch I still hate its interface and it sends emulation back about 20 years but it does have a current core for MAME unsure if this handles touchscreen any better as my device is old?

https://docs.libretro.com/guides/arcade-getting-started/

[johnparker007]

Paul-Arnold has fixed the long-standing regression with meter error on boot on sc4 (and MPU5 not yet working):
https://github.com/mamedev/mame/commit/4479bd973106b42b0cfecbd531101d532631ecb5

[Altharic]

Excellent i seem to recall many layouts are already in mame for both mpu5 and sc4 i should be getting a steam deck shortly so will be able to test 

[Altharic]

https://github.com/mamedev/mame/commit/654dbd8285ca8b569f54fb133782d8f3087a005f

 

layout/bfm_sc5_gu96x8.lay: Use loops and save many lines.

Also various cleanup.

[Altharic]

https://github.com/mamedev/mame/commit/a42a71d5c5bd2cce99bd671a86982ed19b24c4f8

 

mame/bfm: Implement Bell Fruit 96x8 dot matrix display for Scorpion 5. (

#11805) * New driver for BFG 96x8 dot matrix vfd * Add 96x8 dot matrix vfd to Bell Fruit Scorp5

[johnparker007]

11 minutes ago, Altharic said:

https://github.com/mamedev/mame/commit/a42a71d5c5bd2cce99bd671a86982ed19b24c4f8

 

mame/bfm: Implement Bell Fruit 96x8 dot matrix display for Scorpion 5. (

#11805) * New driver for BFG 96x8 dot matrix vfd * Add 96x8 dot matrix vfd to Bell Fruit Scorp5

First CL is just housekeeping, second one is good though - finally get MAME sc4/sc5 on the dot alphas, they've previously been running in 14/16 segment alpha back compatibility mode  

[Altharic]

7 minutes ago, johnparker007 said:

First CL is just housekeeping, second one is good though - finally get MAME sc4/sc5 on the dot alphas, they've previously been running in 14/16 segment alpha back compatibility mode  

Seen the civil war from Haze on X?

I kind of see his point

[Johnnyafc]

4 hours ago, Altharic said:

Seen the civil war from Haze on X?

I kind of see his point

And here 

https://github.com/mamedev/mame/pull/11937

haze seems a decent chap watched few of his live streams etc 

mames going the way that fme did

[johnparker007]

9 hours ago, Altharic said:

Seen the civil war from Haze on X?

I kind of see his point

I've not seen that, though I may scout it out, but I did just read through the github thread from @Johnnyafc and I'm not surprised by what I found.

It is not as simple as it appears; on the one hand, it does appear that Cuavas is a bad guy stopping emulation coders doing their thing.  But on the other hand, MAME has grown into the most complex emulation project in human history by an order of magnitude.  It emulates such an insane amount of hardware, and ties it all together.

In the past, emulation devs were allowed to do their thing, and while that worked ok when it was just video games, however once the scale massively jumped, with fruit machines, home PCs, main frames, synthesizer keyboards, calculators, word proessors, door bells etc etc etc, the list goes on to emcompass anything that can be emulated... it became a real nightmare codebase.

And Cuavas, love him or hate him, is a very intelligent coder with the aptitude to be able to deal with that complexity and try to organise it in some fashion.

But!  I also totally see the other side of it - just the other day Paul (who's been looking at some sc4/5/6 stuff), was saying he doesn't do much, because he hates the insanely strict code reviews.  He doesn't mind doing the reverse engineering (which is the hard bit), but hates trying to then fold his work into MAME's monolithic codebase, so doesn't actually bother.

David alludes to this on the end of his github post:

"You're driving people away, making the project look bad, unapproachable. MAME is getting a reputation as very gatekeepery, and elitist, looking down on anybody who has a slightly lesser understanding or different set of skills."

I have to agree.

Not sure if it would be different with someone else at the helm... or if it is just that MAME has grown so large as to be a very slow moving frustrating project.

In David's case he was just partway through writing a disassembler, and committing a WIP from what I can see.  From Cuavas's perspective he is saying, get it to a state where it passes the strict MAME coding guidelines before committing (mainly; using official MAME assembler syntax, some other bits, plus the infy loop is just an oversight).

I can see both sides of this to be honest.  I'm not sure if there might be a branch in the future 'casual MAME', that would have more relaxed coding rules, the people who are into the rigid code standard stuff could import work from that branch into the core MAME branch... though that it introduces its own issues.

Can definitely agree that is does discourage a lot of emulation coders from even bothering, since they know they then have to deal with the iron fist of the Cuavas code reviews!  But, Cuavas does loads of work, and the code reviews generally are all valid criticisms, compared to the coding standards and methodologies of what should be merged into the root codebase.

I've posted that here, but I'm staying well out of it over there!  

Edit:  I just read though some of David's twitter, and I saw:

It also appears that HE is allowed to go through an iterative process of how the disassembler represents things, through a series of live checkins, as he understands it, but when I want to do that, not allowed
Unless he's finally realising that yes, that iterative process matters

So there may be some level of double-standards here that I wasn't aware of, i.e: Cuavas may not be following his own rules when he is working on something in MAME, or making up arbitrary ones to suit what he is doing at the time... not a simple situation this... I can really see both sides, as it's putting a lot of people off even working on the project

Edited January 24 by johnparker007

[Altharic]

And where are those guidelines........that's the problem it's a case of do as I say not as I do 

[johnparker007]

3 minutes ago, Altharic said:

And where are those guidelines........that's the problem it's a case of do as I say not as I do 

Yes and no... there are most cases where Cuavas is correct, and the code is not 'up to the MAME coding standards' (the basics are here: https://wiki.mamedev.org/index.php/MAME_Coding_Conventions).

But - lots of people who enjoy reverse-engineering stuff, building emulation - don't enjoy being tied to rigid coding standards.  So they just don't bother working on MAME... and I've seen this happen in real time the other week with Paul, he just hates all the coding standard stuff.

I'm not sure what the solution is - one way might be to keep Cuavas, but then have an army of people who enjoy converting code from its current form to one that meets coding standards... I dunno, without Cuavas things get messier again, and the codebase is huge and precise, perhaps more than any other codebase that has ever existed.  I can't think of an equivalent project.

[slotsmagic]

7 minutes ago, johnparker007 said:

Yes and no... there are most cases where Cuavas is correct, and the code is not 'up to the MAME coding standards' (the basics are here: https://wiki.mamedev.org/index.php/MAME_Coding_Conventions).

But - lots of people who enjoy reverse-engineering stuff, building emulation - don't enjoy being tied to rigid coding standards.  So they just don't bother working on MAME... and I've seen this happen in real time the other week with Paul, he just hates all the coding standard stuff.

I'm not sure what the solution is - one way might be to keep Cuavas, but then have an army of people who enjoy converting code from its current form to one that meets coding standards... I dunno, without Cuavas things get messier again, and the codebase is huge and precise, perhaps more than any other codebase that has ever existed.  I can't think of an equivalent project.

I'm not clued up on MAME and all the politics so apologise for this uninformed opinion, but wouldn't the attitude shown by Cuavas just drive people towards their own standalone emulation projects, or possibly Retroarch cores as an example?

[johnparker007]

6 minutes ago, slotsmagic said:

wouldn't the attitude shown by Cuavas just drive people towards their own standalone emulation projects, or possibly Retroarch cores as an example?

Kind of yeah.  A whole emulation project is iffy as it's a massive undertaking - but if someone wanted to say get all of sc4/5/6 fruit machines working, they might make a 'fork' of MAME, to say ScorpionMAME.  Then do all the work in there, free from code reviews.  But ideally you'd want someone to then merge that work back to MAME.  It doesn't need to all start splintering out like that into these dedicated mini-projects, as there'll be loads of duplication of work, bugfixes not making it across the forked projects... it's really not ideal longterm.

Retroarch doesn't really do any emulation development, it's just a 'wrapper' that encapsulates existing emulation projects into its 'core' format.  A lot of devs in the emulation community hate it, though I can see it's utility, it does overstep sometimes with building cores where permission is not granted etc.

[Altharic]

3 hours ago, slotsmagic said:

I'm not clued up on MAME and all the politics so apologise for this uninformed opinion, but wouldn't the attitude shown by Cuavas just drive people towards their own standalone emulation projects, or possibly Retroarch cores as an example?

Retroarch is a backwards step in addition to creating unneeded and out of date cores they also use obsolete emulators solely because they just run  an example of this is handy whilst it runs most Atari Lynx games its emulation dates from the late 90s aside from being ported to SDL its not been worked on since great if you just want to play games but not accurately

 

3 hours ago, johnparker007 said:

Kind of yeah.  A whole emulation project is iffy as it's a massive undertaking - but if someone wanted to say get all of sc4/5/6 fruit machines working, they might make a 'fork' of MAME, to say ScorpionMAME.  Then do all the work in there, free from code reviews.  But ideally you'd want someone to then merge that work back to MAME.  It doesn't need to all start splintering out like that into these dedicated mini-projects, as there'll be loads of duplication of work, bugfixes not making it across the forked projects... it's really not ideal longterm.
 

You mean like Agemame was to begin with before it was swallowed into mame

Edited January 24 by Altharic

[johnparker007]

1 minute ago, Altharic said:

Retroarch is a backwards step in addition to creating unneeded and out of date cores they also use obsolete emulators solely because they just run  an example of this is handy whilst it runs most Atari Lynx games its emulation dates from the late 90s aside from being ported to SDL its not been worked on since great if you just want to play games but not accurately

 

You mean like Agemame was to begin with before it was swallowed into mame

Yeah I don't know what the solution is with MAME currently, it's all kinds of tech that's being impacted (in terms of people don't want to work on them with MAME currently as it is, and these are highly skilled emulation people, who you definitely do want working on this stuff), fruit machines is just a small part of it... going to keep to the sidelines for now!

[Altharic]

On 21/12/2023 at 20:36, Altharic said:

Anyone able to try on a pc with a touchscreen does it pick up the touches if the windows one doesn't the android one won't?

I just tried with a steam deck and touching the buttons does nowt so the same behavior as the android version

Which games have layouts setup for touch?

[johnparker007]

15 minutes ago, Altharic said:

I just tried with a steam deck and touching the buttons does nowt so the same behavior as the android version

Which games have layouts setup for touch?

The rom j6indy is 'completely' set up for touch (some are only set up partially and do not have an orange mouse pointer).
The rom m4andycp is 'partially' set up for touch (doesn't have the orange mouse pointer, but buttons can still be clicked with OS mouse pointer).

[Altharic]

Mame4droid has been updated in line with the current build

 

https://github.com/seleuco/MAME4droid-2024/