CLR.CLS (ooRexx.NET) cannot initialize jni4net 0.8.8 on Java 11+: LinkageError on System.runFinalizersOnExit
Hi Shmuel, yes this is a permanent restriction as the used library breaks with Java's internal changes starting with Java 11, so you need to use Java 8 for this functionality of BSF4ooRexx. This is not as terrible/difficult as it may seem as it easy to have multiple Java versions installed on the same machine. The restriction is permanent as there is a new .NET library for ooRexx 5.3.0 beta in development, which includes CLR.CLS support. It already runs all of the BSF4ooRexx .Net samples and supplies...
CLR.CLS (ooRexx.NET) cannot initialize jni4net 0.8.8 on Java 11+: LinkageError on System.runFinalizersOnExit
20261103 Fix id2x() for 32-bit operating systems
20261024 Committing all files that made it in today's refresh.
rexxj has invalid syntax for W11 directory containg space
Fixed with [r1129].
20261024 Fix [bugs:#56]
20261024 Adding Claude's test units for JDOR from 20261014
JDOR: printPos fails when called without arguments
Fixed with [r1127].
JDOR: drawImage passes width/height instead of second destination corner
Fixed with [r1127].
Fixed with [r1127].
20261024 Applying Claude's patch to fix [bsf4oorexx:bugs#57] and [bsf4oorexx:bugs#58] and newly found by Claude on 20261014
20260323 make sure for levelgetLanguageLevel(), that the native cached Java method IDs got initialized already by creating explicitly a RexxEngine and let it apply some Rexx code
Install does not update environment variables
Update: with the zip unblocked before unzipping and install.cmd run from bsf4oorexx\install\windows, BSF4ooRexx850 (v850-20260326-refresh) installed successfully on Windows 11 with IBM Semeru 21.0.5 (OpenJ9, 64-bit) and ooRexx 5.2.0. A plain rexx script with ::requires "BSF.CLS" now works from a fresh shell. Thank you; fine to close #55. #56 (rexxj.cmd and the unquoted path) still applies.
JDOR: printPos fails when called without arguments
JDOR: drawImage passes width/height instead of second destination corner
Hi Shmuel, it seems that if you un-quarantine per the instructions the install will work. Therefore I am about to close this bug report as invalid.
That sound like the issue; I'll try it. Thanks.
Thank you, Shmuel! According to the readme, before unzipping, you should unblock the zip archive, either via the zip file's property pop-up (mark unblock, then press apply), or use "powershell unblock-file *" Did you unblock before unzipping? If not, it would explain why the installation does not work, as Windows itself will inhibit the execution/usage of blocked programs and DLLs. The readme states to either change into bsf4oorexx\install\windows and run the script manually from there, or to double-click...
BSF4ooRexx install/environment status — summary for Rony and the ticket System: Windows 10.0.26200.8875. OpenJDK 21.0.5.11 (IBM Semeru, Eclipse OpenJ9), confirmed same bitness as ooRexx (both 64-bit). Install, in order actually run Downloaded the BSF4ooRexx installer zip, unpacked to a staging folder under Downloads. First attempt: ran install.cmd by full path from C:\Users\Owner (i.e. not the install directory) — failed immediately: Error 3.901: Failure during initialization: Program "kickoff.rex"...
rexxj has invalid syntax for W11 directory containg space
Install does not update environment variables
IA 32-bit on AMD 64-bit platform BSF4ooRexx850
DialogBox Return codes not as described
Hi Jon, unfortunately the slide's information on foil 11 is incomplete. bsf.dialog uses javax.swing.JOptionPane (https://docs.oracle.com/javase/8/docs/api/javax/swing/JOptionPane.html). In the case that the fourth option is present (implies argument # 6 blank delimited string of buttons, and argument # 7 default button are both not supplied), then one can supply one of the default buttons: default (only ok button gets displayed) yesNo yesNoCancel okCancel Then the return values are defined to be:...
DialogBox Return codes not as described
Jan-Erick, this indicates that in the mix of rexx.exe, jvm.dll and BSF4ooRexx there is a bitness mismatch. All components need to either be of 32- or 64-bits. Also, if you intend to use AOO on Windows, be aware that AOO on Windows is currently available in 32-bit only, such that you cannot interact with AOO from 64-bit programs. What is the bitness of rexx (rexx -v) and of java.exe (java -version) that you use in this invocation (you need to look up rexxj.cmd to know for sure)? If you install or...
Hmm, 641 gave a similar message that it wouldn't run, but in the end the script (instrx.rex) has done something so that one can use BSFooRexx 641. But the problem remain with 850, that I can't overcome.
IA 32-bit on AMD 64-bit platform BSF4ooRexx850
AOO on Windows is 32-bit, net-oo-rexx is 64-bit. You could replace the ooRexx part in net-oo-rexx with the portable 32-bit version of ooRexx by copying it to net-oo-rexx/oorexx. In case you want more information about Java exceptions you could do the following: add a "signal on syntax" before the statement where exceptions occur, create an internal procedure "syntax:" and add debug code, e.g.: signal on syntax ... code ... exit syntax: co=condition('object') say ppJavaExceptionChain(co) exit -1 However,...
Can't see that portable would help me as my scripts call 32-bit and 64-bit apps right and left to use whatever resources present on the different computers it encounter. What does this message mean and need to do under the hood? Is it a IPC pipe it refer to that is the trouble? ---> com.sun.star.lib.connections.pipe.PipeConnection Something one can mend here? Error 40 running C:\Program Files\BSF4ooRexx850\UNO.CLS line 1858: Incorrect call to routine. Error 40.900: BSF4ooRexx850/routine/BSF(), error...
20260117 Refresh version.
Hello Rony, I hope that I am the bearer of good news … It took me a bit of time to find the download site for Packages and I found : http://s.sudre.free.fr/Software/Packages/about.html I hope that it is the right one After that I expanded ( it took a bit of research ) and listed for all the executables the architecture ( lipo -info ) And all of them report : Architectures in the fat file: <filename> are: x86_64 arm64 The Packages application is GOOD,</filename> Did the same for the …BSF4ooRexx.....
Hello Rony , Unfortunately opening the .pkg file does not even start the installation, I get immediately a pop up asking to install rosetta, so the installer program might be the culprit Checking the architectures of the executables in the pkg was my first idea, But navigating thru the pkg with pkgutil to extract the payload is NOT easy I will keep you posted Bye for now :-) Enrico On 13 Jan 2026, at 18:35, Rony G. Flatscher orexx@users.sourceforge.net wrote: I am speechless too! It is a universal...
Hi Enrico, I am speechless too! It is a universal package just run "file" on the binaries and you will see that it got compiled for both architectures, Intel and ARM (M?). Maybe the installer software that I use ("Packages") is Intel only and triggers Rosetta2? Another possibility is that you have the Intel version of Java on board (this is a JOKE!) ... :) ;) Seriously, if you can find a way to not trigger the need for Rosetta I would highly appreciate it, not being quite fluent on Apple. (Also,...
Universal BSF4ooRexx850 with 64-bit ooRexx5.2.0beta (r13065) 20260112.pkg is NOT universal
20260112 save current state
20250112 lsregister's kill option got removed without notice and causes an error in the postflight scripts. Removing that option makes the GUI installation work again.
20260108 Fix rexxjh.cmd to be only verbose if BSF4Rexx_quiet<>1.
20260106 Update the MacOS package version.
20260106 Add additional debug output, point to rexxj2 (not rexxj).
20260106 Update BSF.CLS, rexxb4r.rex, add OpenJDK/Java 25 System property html listing.