EMULAB Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

The new forum is online, hope you enjoy it!

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Roman

Pages: [1] 2 3 4 5 6 ... 167
1
ok...unneededMasks are in....you can specify a regular expression to filter unneeded files/folders in rompaths and in the set's archives/folders. So to filter out ".DS_Store" on any location you could enter something like ".*\.DS_Store$"

(any char or no char before .DS_Store and the . is escaped with \ and .DS_Store is at the end of the name ($))
Surely you can create a regular expression which matches more one specific ending/whatever, e.g. by using |

2
clrmame Discussion / Forum Spam Bot Protection
« on: 07 July 2024, 19:47 »
Looks like the forum reacts a bit weirdly and more strictly to posts today....I also had problems getting a reply through....used a different browser and it worked though.
Already informed the forum owner about it...maybe some software update causes this...

3
clrmame Discussion / Re: MAME .267 and the new scanner
« on: 07 July 2024, 19:42 »
Normal for MAME
also see readme.md

"When loading an input file you might see some warnings. For a standard MAME
-listxml you e.g. see sample specific warnings. It's mainly about sample
relationships from machines to a sample parent machine which is not available
in the XML. Such sample-only sets are generated automatically so that the
assignment is correct again. Similar warnings exist for the use of samples
which aren't available in the sample parent set. This is also fixed internally."

You can ignore them.

cmpro also complains about them (see attached screenshot)

4
clrmame Discussion / Re: MAME .267 and the new scanner
« on: 05 July 2024, 19:45 »
little peak for a possible new release somewhere next week...

added: ui: log and tree can be vertically resized individually now (split move possiblity between those 2 controls)
added: core: support more XML based datfiles
added: core: filter option supports prefix 'file:' to specify a file holding setnames used for filtering
fixed:   core: several obsolete "can't remove" errors based on not taking an in between rename into account
fixed:   core: rom/disk/sample names may contain illegal path characterss (incl. ending with .)
fixed:   core: fix-missing can skip one file from a batch and doesn't fix or report it


I also have some things on my wish list (unneeded masks, disable merge box when no parent/clone relationships, etc) which I also want to add.....after that I want to spend time on the profiler and bring scanner/rebuilder together in one app...so if you have something tiny and useful, speak now ;-)

5
I put it on the wish list...maybe in a first run you'd need to add it somewhere in the settings.xml ...time will tell....

6
oh I was only talking about the new scanner :-) Good to hear they show up fine there.
Yes it's faster than old cmpro due to various reasons. Threading, completely different approach to handle merge modes, etc etc...

Actually I don't want to spend time on old cmpro anymore unless something breaks....

7
Regarding speed: well...what do expect when scanning over the network :-)


Update: hmm...it shows mapped network drives for me in the left part of the folder and file dialogs and they are selectable

8
hmm..haven't tested it yet, but surely you can simply enter filepaths in UNC syntax like

\\fileserver2\MAME\roms

9
clrmame Discussion / Re: MAME .267 and the new scanner
« on: 01 July 2024, 06:20 »
ok...little update (no binary yet though)

- fixed the handling of illegal file characters as rom/disk/sample name, including a "." at the end. This is now aligned to old cmpro and the gtsers7a issue is gone

- some fixable but missing files don't appear or get fixed in a first scan run (cmv4zga etc...), found the issue...lousy "off by 1", it simply forgets to process one entry during a fix missing operation...


Will check some more things before I put out something new.....

10
Hmm...yep, that's a problem.
Since we talk about uncompressed chds, you can't turn on the hashcollision mode (new scanner/rebuilder would do that automatically) which would use subfolders for each clone in a full merged environment. Why? Because MAME does not load decompressed files in subfolders, otherwise it would work.

Either use split merged mode or report this to MAME developers so it gets somehow fixed.

In the meantime I think about something which may be possible as a workaround.



I've reported it in the MAME discord tester channel already...

11
clrmame Discussion / MAME .267 and the new scanner
« on: 30 June 2024, 12:19 »
So new MAME is finally out....time to check the scanner on it... and yes, there are some things I need to look into.

Trivia (Questions Series 7, alt question ROM) [folder: gtsers7a - parent: gtsers1]
Here either cmpro or the new scanner complains about the name. The rom name ends with a "." which is not possible under Windows...cmpro removes it from the name, the new scanner seems to keep it, so you end in a rename loop if you scan with cmpro then with the new scanner and so on...I will fix that

cmpro also complains about some missing clone roms in cmv4zga, gtsers7a ,mspacmane , mspacmane2, pleiadsgmp and pacmanmr while the new scanner doesn't do that in its first run. So you may have the feeling everything is ok, while a 2nd scan reveals the "fixable but missing" ones. They also get fixed, but they should get fixed in the first run already OR it should be listed.

I will definetly check on those and what the reason for them is....so, if you find other things besides of these sets, let me know

12
Windows does not allow (or actually auto removes) "." at the end of a filename, that's why such files get the . removed when parsing the datfile. You can have it in an archive though.
Hmm..both apps complain about the file when "." is at the end....I will check it. Correct would be without "." at the end since Windows removes it on itself.....I will do some more test on it.

Regarding your requests:

- yes, resizable logging versus resizeable scan tree (i.e. being able to resize both with a split element or something) is on my list
- yeah, for now, the combo boxes simply keep the last 5 unique entries...currently you can only clear them in the settings.xml, good point

13
ok, I've put it on the wishlist....should be pretty easy to add

14
what exactly do you mean with "import set names from text files"?

You mean you have a textfile holding set names (either line by line or somehow separated by a character?) and you want to use that as "Filter". So if your file holds "pacman" and "005" you want that the loaded datfile is limited to these two (plus their dependencies).

If that's the case, I could think about enhancing the filter (which currently does regex and xpath) maybe by entering something like "file:c:\temp\list.txt".
Or do you mean something else?

15
4.015 (2014-08-19)
....
    misc: parse rom merge tags enabled by default (on clean install)

Over-a-decade-users may have that ;-)

16
Check your profiler->options->Parse ROM (and DISK) 'merge' tags.

Both should be enabled (which is the default) and I think you got them disabled which would lead to various additional (hash identical) needed files .

This is also mentioned here: https://www.emulab.it/forum/index.php?topic=9767.0

17
clrmame Discussion / New scanner versus old cmpro
« on: 24 June 2024, 16:28 »
Trying to collect stuff here which may be helpful when you switch or play around with the new tools:

Differences in Merge Modes:
- split: this should be fully identical to cmpro
- standalone (new) versus not or unmerged (old): The new standalone mode is more accurate in my opinion since it includes all dependency files (parent/bios/devices) into the set. The old mode only takes over the parent ones (and with some tricks the  bios)
- full merged mode: the full merge mode acts like cmpro's full merge mode with "full merge mode -> hash collision name" (this is not the default options) enabled. The naming there is a bit misleading. It simply means that all clone sets inside a packed parent set are stored in subfolders.


Regarding scan results there can be slight differences:
- the tree is finally sorted alphabetically be machine description, cmpro is by setname which leads to some weird output order
- some results might slighly differ since the new scanner is simply different and the categorization of a wrong named machine or a wrong placed machine or an uneeded machine differs a bit. But after fixing, the resulting sets should be the same.

Scanning software lists or using system default paths
There are no system default paths anymore and you don't need to go through rompath setup hell when it comes to software lists collections.
Simply us 1 rompath and use either one of the patterns (e.g. #BIOSSPLIT# which is identical to system default paths) or in case of scanning a -listsoftware collection, you can even keep that blank since it internally uses #SOFTLIST#. You only need to store your software lists in rompath subfolders named after the list though.
For example you use: rompath: c:\mame\softwarelist and inside there are folders like a2600, a2600_cass, c64_flop_misc, amiga_hardware and so on which hold the files for the belonging lists.


Parent/Clone relationships
In cmpro it it optional to use merge tags for roms/disks from the datfile to identify parent/clone relationships of rom files. The new applications fully rely on it (at least for non software lists).


Supported Datfile Formats
Old cmpro supports all kind of datfiles. New apps only support XML based ones which are more or less compatible to -listxml, -listsoftware or MAME's hashfile format. This doesn't mean you are limited to MAME.

18
Hi,

the answer is: it depends :-)

There are some main differences in the merge modes:

- standalone (new) versus not or unmerged (old): The new standalone mode is more accurate in my opinion since it includes all dependency files (parent/bios/devices) into the set. The old mode only takes over the parent ones (and with some tricks the  bios)

- full merged mode: the full merge mode acts like cmpro's full merge mode with "full merge mode -> hash collision name" enabled. The naming there is a bit misleading. It simply means that all clone sets inside a packed parent set are stored in subfolders.

- split: this should be fully identical to cmpro


Regarding scan results there can be slight differences:

- the tree is finally sorted alphabetically be machine description, cmpro is by setname which leads to some weird output order
- some results might slighly differ since the new scanner is simply different and the categorization of a wrong named machine or a wrong placed machine or an uneeded machine differs a bit. But after fixing, the resulting sets should be fine.


I assume you're using full merged sets and don't use the "keep clones in subfolders" mode in cmpro.....Personally I think it's a bit clearer to store the clones in subfolders (however that only really works when you keep the archives zipped/7zipped).


I'm wrong with my assumption about the full merge mode, please let me know an example of a set which is differently handled.

19
Well, if you want to play around with the tools outside of the MAME world (i.e. load other datfiles which clone the XML layout from MAME (e.g. dir2dat created ones and others floating around), you can use

https://mamedev.emulab.it/clrmamepro/binaries/scanner0011_rebuilder0081.zip

This allows you to load more datfiles than the typical MAME ones

20
Finally....

I've decided to put out the first instance of the new scanner (bundled with a rebuilder update).

If you followed the project here on the forum in the past, you'd know that this is not yet a full clrmamepro replacement. It's something I had fun with by rewriting things in a way clearer and more modern code style. If you want to know more about the benefits, you will find information in the readme file. There is no profiler yet (that's the next thing to do) but you are already able to switch between a handful of setups (e.g. a -listxml and -listsoftware profile) which actually should be enough for a MAME user. Regarding the UI versions, they are preliminary. Things might change a little or a lot. The goal of these releases is to get robust cores. I highly recommend to have a peak at the readme file and of course the usage (especially turning on fix options) is your very own risk.

If you're enjoying this little project as much as I do, feel free to support me

Pages: [1] 2 3 4 5 6 ... 167

Page created in 0.184 seconds with 20 queries.