[quote] ... It contains important bug fixes. ... OS3 and AROS ports of smb2-handler are now also available at:[/quote] Great ! Thank you very much for the update :thumbsup:

Version 54.3 of the OS3/AROS filesysbox.library is now available at: https://github.com/salass00/filesysbox/releases It contains important bug fixes. OS3 and AROS ports of smb2-handler are now also available at: https://github.com/salass00/smb2fs/releases As Zune didn't have an equivalent class that I could use to replace the ClassAct/ReAction requester.class (used for the password requester) I simply disabled this code in the AROS port for now, so on AROS the password will have to be specified in the DOSDriver/Mountlist entry. @magorium You're welcome.

Thank you very much for the update salass00 (y)

Just over seven years after the release of version 53.6 (looks like I never created a binary archive for V53.7) version 54.1 is now available for i386-aros (and also m68k-amigaos) at: https://github.com/salass00/filesysbox/releases The V54.0 version which brought the AROS/OS3 backport up to a similar level as the AmigaOS 4 port feature wise was thanks to the work of BSzili. Changes since the previous release are: filesysbox.library 53.7 (26.4.2018) - Got rid of all the Get****Base macros. - Fixed an out of bounds array access in FbxUpdatePaths() which would result in memory after the FbxFS struct being trashed. filesysbox.library 54.0 (22.12.2022) - Added timezone conversion code using locale.library. - Calculate a path hash for fib_DiskKey unless FBXF_USE_INO flag is set. - Added FBXF_USE_FILL_DIR_STAT flag to avoid unnecessary getattr() calls for filesystems like ssh2-handler that already provide all the stat data in readdir(). - Don't treat ENOSYS return value from FbxSetAmigaProtectionFlags() as an error. - Rewrote the FbxFillInfoData() function and made it set id_BytesPerBlock to 512 instead of zero if a volume is not available. - Modified path handling for Ext2FileSystem which only accepts "/" as path argument for the file system root directory and not "". - Rewrote FbxReturnMountMsg() using ReplyPkt() and added a NULL msg safety check. - Added some more validity checks when reading FSSM from device node in FbxSetupFS(). - Added validity checks in code for getting FSSM in FbxQueryMountMsg(). - Changed CHECKVOLUME() macro to return ERROR_NOT_A_DOS_DISK instead of ERROR_OBJECT_IN_USE when the filesystem is inhibited. - Added semaphore protection in FbxHandleTimerEvent(). - Added the FbxQueryFS() and FbxQueryFSTags() functions. - Added the FbxGetSysTime() function. filesysbox.library 54.1 (14.1.2023) - A rename operation with exactly the same source and destination is now treated as a no-op and returns success rather than ERROR_OBJECT_EXISTS. - Rename operations where the the only change is in letter casing (e.g. "rename tmp TMP" ) no longer fail with ERROR_OBJECT_EXISTS on case insensitive file system implementations. - Advance the file position in FSWrite() also if write() returns a lower number of bytes written than was specified. - Implemented the same charset conversion system that is used in the latest AmigaOS 4 filesysbox versions. Internally filesysbox still uses UTF-8 for all strings and conversion only happens when strings are passed to or from the AmigaDOS API. In order for the conversion to be reversible any UTF-8 characters that do not have an equivalent in the local charset will be converted into URL-style escape sequences where the character is specified as a base32 encoded character sequence preceded by a percent symbol.

Happy New Year !!

Happy New Year everyone!

That's fantastic work and thank you for all that! I agree with you, it was time to bring Janus-UAE up to date. I believe if we can manage to port SDL 2.0 to AROS, it is going to open a gate to a lot of new applications, games, etc. etc. but one thing that I noticed is the OS is still somewhat very unstable. I recently managed to get AROS One 1.8 by Amigasystem on AROS-EXEC working (Supported Video, Wifi, and Audio) on a Dell Latitude D520. I struggled with the Wifi card but figured out what was the problem (will post an article in here soon). I am still experiencing occasional lock-ups, unexpected crashes, and annoyances such as a shell icon on the desktop showing multiple times after placing just one via "leave out" option. It's frustrating and distracting when I am trying to focus on one particular problem/task and other things around go crashing. In any case, it is still fun in trying to port a piece of software for the first time on a new OS like AROS :)

[quote name=amigamia post=2117]@amigamia - [quote name=MiDWaN post=2116]@MiDWaN - Was there any more progress on this matter? I would like to see if I can improve Janus-UAE on AROS, and one of the things would be moving it to SDL2 - but that requires the library first, of course :)[/quote] Hello MiDWaN, welcome to AROSWorld! I haven't made much progress unfortunately. It's on pause right now as I had to redo my virtual machines hosting AROS on them. However, I plan on picking it up again but I am no expert. I am doing it it as a learning experience as well as because no one else seems interested on porting SDL 2 to AROS. I thought I would give it a shot. I was in contact with one of the AOS4 developers who suggested to use the MorphOS code because it's closer to AROS. I have the code here on this web site but I am going to put it on GitHub soon. Would you be interested in collaborating?[/quote] Hi there! No problem, I'd love to collaborate on this - I would probably have to look into it myself otherwise. I just managed to get everything to compile (using SDL1) yesterday, after a lot of trial and error. The sources of Janus-UAE were a little stale for a few years, and things had changed in the meantime. That means that the "fun" part is only starting, as the first requirement (and the pain in the butt) was making sure I have an environment that I can use to test new builds in (AROS itself, and a fool-proof way of building Janus-UAE). Well, now I have more than one (Ubuntu under Windows 11 with WSL2, Manjaro Linux) and I know more about the pitfalls of the process. :) In other words, no rush and no expectations from my side. Happy to help wherever I can! My work on Janus-UAE is on Github, since I've made some modifications already (with plenty more expected to follow): https://github.com/BlitterStudio/janus-uae

[quote name=MiDWaN post=2116]@MiDWaN - Was there any more progress on this matter? I would like to see if I can improve Janus-UAE on AROS, and one of the things would be moving it to SDL2 - but that requires the library first, of course :)[/quote] Hello MiDWaN, welcome to AROSWorld! I haven't made much progress unfortunately. It's on pause right now as I had to redo my virtual machines hosting AROS on them. However, I plan on picking it up again but I am no expert. I am doing it it as a learning experience as well as because no one else seems interested on porting SDL 2 to AROS. I thought I would give it a shot. I was in contact with one of the AOS4 developers who suggested to use the MorphOS code because it's closer to AROS. I have the code here on this web site but I am going to put it on GitHub soon. Would you be interested in collaborating?

Was there any more progress on this matter? I would like to see if I can improve Janus-UAE on AROS, and one of the things would be moving it to SDL2 - but that requires the library first, of course :)

Sorry to go out of topic but are there any news about PortablE development?

I still have a copy of AmiDevCpp, but I am looking for a Linux-based cross-compiler, and really one that can be scripted easily.

Hi Chris, Good to see you around again. I am not sure if that is available anymore but I do have a copy of the old AmidevCPP which was the full IDE with Amiga and AROS libraries for cross compiling but from Windows. I can send you that if you are interested.

Does anyone know where I can find a version of GCC/G++ compiler for Linux that will cross-compile for AROS? ZeroHero used to provide one, but his website is down, and no-one seems to have saved his AROS (nor MorphOS) versions: http://www.zerohero.se/cross/aros.html (no-longer exists)

Just to make sure we are on the same page: You are using the source-code that is found at the github link you posted earlier ? I'm puzzled about the real suffix..... Unless i overlooked i did not see one of the macro's expand to add such suffix. Note that creating a dynamic version of the library is something else then creating a static one. Trying to compile as the library as linux library isn't the way to go there. You are better of trying to see what is inside the amigaos4 makefile and try to replicate that (minus the ppc specific parts ofc). I wish i had more time on my hands so i could give it a shot myself. Cavemann/deadwood (or anyone else with a bit of c/make knowledge) not around ?

I issued the [code]nm libSDL2.a > lib.txt[/code] and I cannot find just SDL_Init for example. I can find SDL_Init_REAL which is strange but maybe that'll work??

Found online that I can use [code]make -j$(nproc)[/code] to bypass the issue with the platform but goes into all sorts of other errors (See attached). So, I decided to give the Makefile.minimal a shot and that one worked after some modifications. Generates the LibSDL.a

Well I tried different things. What worked first was to run a slightly modified Makefile.minimal to match the correct Includes and gcc [code]make -f Makefile.minimal[/code] This generated a libSDL.a file but I get the same issue when compiling against this library. I also was able to run the configure by tricking it into thinking it's a linux box with the following command: [code]./configure --build=i386 --host=i386-unknwon-linux --disable-video-opengl --disable-video-x11 --disable-video-rpi --disable-pulseaudio --disable-esd[/code] This goes all the way to the end. In the process it asks for an empty volume and an opt as well but clicking on cancel makes it continue. See the attached for the end result of that. This generates the default Makefile and if I issue the following command with no switches, it reads the Makefile that was just generated [code]make[/code] However, it complains about the platform

Well, the routines as mentioned in your screenshot are definitely part of SDL.c. Have you tried the nm tool (might have full platform arch included in the name) to see what symbol-names are exported in your SDL.a archive ? Your test-program tries to locate the symbols in the SDL library that you've created but seem unable to find/locate them. There can be a bit of gcc voodoo involved (e.g. parameters/options that I don't know about) to get those symbols visible as intended. Have you followed the amigaos4 makefile when you created/build your SDL lib ? (as that probably is your best bet with regards to 'being similar to AROS').

Hmmm, I have the libSDL2.a that compiled earlier and placed it in /Development/lib. I also created a new folder called SDL2 in /Development/include/SDL2 and copied the .h files from the source archive I created a Makefile according to [url]https://lazyfoo.net/tutorials/SDL/01_hello_SDL/linux/cli/index.php[/url] I know it's finding the correct library (libSDL2.a) because if I rename it I get a different error. However, it can seem to find the SDL_Quit. SDL_Init, and a bunch of other SDL_****xx functions. (See screenshot) The source file 01_hello_SDL.cpp has been modified to include the SDL2/SDL.h Any idea?