bugGNU Octave - Bugs: bug #63971, Invoking Octave's Java subsystem...

 
 

bug #63971: Invoking Octave's Java subsystem crashes Octave

Submitter:  None
Submitted:  Tue 28 Mar 2023 10:37:48 PM UTC
   
 
Category:  Octave Package Severity:  3 - Normal
Priority:  5 - Normal Item Group:  Installation Failure
Status:  Invalid / Not an Octave Bug Assigned to:  None
Originator Name:  Originator Email:  -email is unavailable-
Open/Closed:  * Closed Release:  * 8.1.0
Operating System:  * Microsoft Windows Fixed Release:  None
Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Jump to the original submission

Wed 10 Apr 2024 08:15:14 PM UTC, comment #24: 

Thanks Dmitri.

Interpretation of the user reports below shows that Octave itself runs fine until the Java subsystem gets accessed (i.e., Octave attaches to a JVM), and then it crashes.

What the superuser.com article that you refer to describes about dll search paths, sounds like it could well be implemented and set as an optional security policy. I.e., more or less what I wrote in comment #22.
BTW my Windows 10 and 11 systems are fully up-to-date so should/could have such restricted dll search options OOTB. Yet I don't see any of that, so I think it's all due to local settings by company IT departments.

That said, maybe the code that Octave uses to attach to or initialize a JVM isn't sufficiently robust to gracefully catch unexpected JVM initialization issues. However I'm insufficiently proficient in C++ and Java stuff to appreciate the relevant stanzas in ov-java.cc and ov-java.h.

Philip Nienhuis <philipnienhuis>
Group Member
Wed 10 Apr 2024 06:23:02 PM UTC, comment #23: 

This is perhaps more suitable for discourse discussion, but googling around I found:
https://superuser.com/questions/133746/how-does-the-compatibility-mode-in-windows-work


There have also been some other changes from Windows version to Windows version. In older versions for example, if a program loaded a DLL, the search path for the DLL also included the current directory. This is a security issue, so newer versions of Windows by default don't search in the current directory. With the proper shim you can simulate the old behaviour.


I have an impression that octave on Windows relies on that (loading dll from current dir).

Dmitri.
--

Dmitri A. Sergatskov <dasergatskov>
Wed 10 Apr 2024 06:15:20 PM UTC, comment #22: 

Thanks for reporting back your solution :-)

Rather than feeling "I've wasted my time" (I don't) I am glad you got it solved. And a nice thing is that you've handed us another trick to try if Octave on Windows users report hard-to-track-down issues here - thank you for that!

That said, it's kind of weird that you need to set Windows 7 compatibility for Octave. I still can' t help to suspect this is related to e.g., security policies silently set by your IT department.
On my (virtual) Windows 10 machine at work (semi-gov, so locked down quite a bit) I'm running bleeding edge dev Octave versions all the time. I never had to set compatibility options for any SW I use - and we have a quite a mix of brand new, up to some really old SW running.

Closing this report with "Invalid / not an Octave bug" as there's no other matching criterion :-)

Philip Nienhuis <philipnienhuis>
Group Member
Wed 10 Apr 2024 08:17:00 AM UTC, comment #21: 

PS. I always forgot to mention that loading the ltfat package produced the same behavior as io

Giorgio B <boomer>
Wed 10 Apr 2024 08:11:46 AM UTC, comment #20: 

Hi, yes, I followed the steps in comment #17, but finally, after some unsuccesful attempts (reinstalling with admin privileges and in a different folder D:\octave), the solution was setting:

"D:\OCTAVE\Octave-9.1.0\mingw64\bin\octave-gui.exe" properties -> compatibility -> run this program in compatibility mode for -> windows 7.

Now upon loading the "io" package, I no longer experience any crashes and the xlsread instruction works smoothly.
I feel kind of stupid that I didn't think of this sooner and made you waste all this time, sorry for that.
Now the only side effect is that savepath no longer updates the .octaverc file, but I will look for the solution elsewhere.


BTW actually windows logs system errors and makes them available in the windows event viewer. Here I found:

Name of the application that generated the error: octave-gui.exe, version 0.0.0.0, timestamp 0x65f3421d
Name of the module that generated the error: ntdll.dll, version 10.0.22621.3235, timestamp ax2c4352c
Exception code: 0xc0000374
Error offset 0x000000000010c1a9
...

Groeten

Giorgio B <boomer>
Tue 09 Apr 2024 03:00:04 PM UTC, comment #19: 


> xlsread(MyXlsFilePath,1,"A6:Z99999","oct")

Yes, but did you edit PKG_ADD so that __init_io__.m doesn't get run? (and restart Octave then?)
AFAICS the issue is related to the Java subsystem, and bypassing __init_io__.m and not using any Java-related stuff will avoid Java.

BTW for a next io package release I'm pondering about avoiding automatically searching and loading Java "libraries" (.jar files) for spreadsheet I/O. Which in turn will shift that burden to users; that's an unavoidable consequence.
There are more reasons for this than just the Java issues underlying thus bug report.
Another big motive for me as io package maintainer is that upstream, a.o., Apache POI, in each new release all kinds of required Java methods are shifted hence and forth between separate .jar files and it's getting too much work for me to keep track of what newly required method is hidden where. In fact I've already given up as far as Java POI is concerned.

Philip Nienhuis <philipnienhuis>
Group Member
Tue 09 Apr 2024 02:06:58 PM UTC, comment #18: 

In my case(*), after following comment #17 steps, the command:

xlsread(MyXlsFilePath,1,"A6:Z99999","oct")

keeps octave crashing.
Speaking with my system administrator, he told me that during SW installation some folders are write-protected.
We'll try to reinstall Octave after removing all restriction and antivirus. I'll keep you informed.

(*) Windows 11 Pro - version 23H2 - Build OS 22631.3296
Installation Date: ‎05/‎04/‎2023
Windows Feature Experience Pack 1000.22687.1000.0

Giorgio B <boomer>
Tue 09 Apr 2024 11:45:18 AM UTC, comment #17: 

@Anonymous comment #16:

What happens when you enter the command
javachk ("jvm")
in the command window?

Did your Java environment change recently?
If your script has been working flawlessly for months but is now breaking things, something must have changed.

BTW I thought that I had entered another comment here but I don´t see it. So I repeat the steps outlined in it:

Rather than renaming __init_io__.m , edit the PKG_ADD file in the io package directory so that the only line it contains is commented out and reads like this:

## __init_io__ ()

and save it back.
Restart Octave, and try if you can use xlsread and xlswrite with always specifying the "oct" interface.

Philip Nienhuis <philipnienhuis>
Group Member
Tue 09 Apr 2024 09:44:22 AM UTC, comment #16: 

Good morning, I would like to point out that I too have been experiencing the same problem from a few days/weeks.

The PC is still the same, I haven't made any changes to the octave (manual update or similar)

The script that has been working for months now causes octave to shut down without any message error.

The line that introduces the problem is

"pkg load io"

On another PC it works fine

I tried reinstalling octave even with a previous version but the problem continues to exist

p.s.
I don't have the skills to support you in finding the bug.
I wanted to write to you to indicate that it is not a problem that only the person who opened the post had

Anonymous
Thu 04 Apr 2024 01:51:28 PM UTC, comment #15: 

Sorry , I didn't notice the double hyphen
In the meantime I upgraded jre, now the ansvers are:

from within octave-->

>> system ("java -version")
java version "22" 2024-03-19
Java(TM) SE Runtime Environment (build 22+36-2370)
Java HotSpot(TM) 64-Bit Server VM (build 22+36-2370, mixed mode, sharing)
ans = 0
>>

from command prompt-->

C:\Windows\SysWOW64>java -version
java version "22" 2024-03-19
Java(TM) SE Runtime Environment (build 22+36-2370)
Java HotSpot(TM) 64-Bit Server VM (build 22+36-2370, mixed mode, _sharing)

C:\Windows\SysWOW64>

So after:
1. renaming '__init_io__.m'
2. restarting Octave
results are:

>> cd (strrep (which ("oct2xls"), "oct2xls.m", ""))
>> pkg load io
error: '__init_io_' undefined near line 1, column 1_
error: called from
    C:\Program Files\GNU Octave\Octave-9.1.0\mingw64\share\octave\packages\io-2.6.4\PKG_ADD at line 1 column 1
    load_packages_and_dependencies at line 56 column 5
    load_packages at line 53 column 3
    pkg at line 639 column 7
>> pkg list io
Package Name  | Version | Installation directory
--------------+---------+-----------------------
          io *|   2.6.4 | C:\Program Files\GNU Octave\Octave-9.1.0\mingw64\share\octave\packages\io-2.6.4
>>
>> io_testscript ("oct", "io-test.xlsx")

Testing interface OCT using file io-test.xlsx...

1. Initialize arrays...
2. Insert first empty sheet...
'col2num' undefined near line 49, column 10

ans =
{
  [1,1] = OCT
  [1,2] = -
  [1,3] = -
  [1,4] = -
  [1,5] = -
  [1,6] = -
  [1,7] = -
  [1,8] = -
  [1,9] = -
  [1,10] = -
  [1,11] = .xlsx
}

I also tried the xlsread function (actually the one I'm interested in), resulting in Octave still crashing at the javaMethod call:

xlsread           line 379 --> xls = xlsopen (fn, 0, reqintf, verbose);
xlsopen           line 308 --> interfaces = getinterfaces (interfaces, verbose);
getinterfaces     line 109 --> [tmp1, jcp] = __chk_java_sprt_ ();_
_chk_java_sprt_ line  26 --> if (! isempty (javachk ("jvm")))
javachk           line 107 --> dum = methods ("java.lang.Runtime");
methods           line  72 --> mtds_str = javaMethod ("getMethods", "org.octave.ClassHelper", obj);

wheras obj = "java.lang.Runtime"

Finally, I could not locate the javaclasspath.txt file anywhere. Is it normal?

Giorgio B <boomer>
Fri 29 Mar 2024 08:40:01 PM UTC, comment #14: 

Thanks for reporting back.
The result of a simple javachk function call makes it clear that it's Octave's Java subsystem that somehow got crippled and the io package isn't the cause (using it just uncovers the bug).
I'll change the bug title accordingly.

BTW please try the command

>> system ("java --version")

rather with one hyphen, i.e.,

>> system ("java -version")

and report back.

The exact Java version doesn't make much difference, on one of my boxes I have a much older version running fine with Octave.
I tend to suspect that your IT dept has done something (policies or the like) that inhibits Octave from properly invoking the Java subsystem. After all the command

>> system ("java --version")

returns an error message emitted by Java (here Octave just invokes a Windows command outside its own interpreter) which means that Java itself is working.

Then:
You didn't report back the results of this Octave command:

io_testscript ("oct", "io-test.xlsx")

(after having disabled the __io_init__ function as per comment #3.
Would you please try that command?
If that works w/o the crash, it (= using the native "OCT" spreadsheet I/O interface) would be a workaround for you.

Another thing that might help is to fiddle with the Java virtual machine (JVM) settings. We have several related bug reports on Java and IIRC one or more pointed to a problem with resetting the JVM memory to some unworkable setting. But I'll have to dig this up.

Philip Nienhuis <philipnienhuis>
Group Member
Thu 28 Mar 2024 09:26:00 AM UTC, comment #13: 

Here are the test suggested in comment #12

OCTAVE COMMAND WINDOW-->

>> system ("java --version")

Unrecognized option: --version
Error: Could not create the Java Virtual Machine.
Error: A fatal exception has occurred. Program will exit.
ans = 1

>>


javachk ("jvm")-> causes Octave to instantly crash

again debugging function 'javachk' the crash occurrs in 'method.m' line 72:
mtds_str = javaMethod ("getMethods", "org.octave.ClassHelper", obj);


COMMAND PROMPT (C:\Windows\SysWOW64\cmd.exe) -->

C:\>echo %PROCESSOR_ARCHITECTURE%
x86

C:\>java -version
java version "1.8.0_301"
Java(TM) SE Runtime Environment (build 1.8.0_301-b09)
Java HotSpot(TM) 64-Bit Server VM (build 25.301-b09, mixed mode)

C:\>

Same result arises from cmd.exe x64.
according to our IT support, java ugrade has been disabled for compatibility with other programs (CAD) I am not a user of, therefore i can upgrade it manually.
Thank you for your efforts

Giorgio B <boomer>
Wed 27 Mar 2024 07:34:34 PM UTC, comment #12: 

So, based on your debug steps in STEP #3 (very good!! thanks for that), it looks like Java might be the culprit.
I don't think, nor can I imagine, that the crash is due to some bug in the io package but rather to some things in your PC.

What is the output of the following Octave commands:

javachk ("jvm")

and

system ("java --version")


Or if Octave crashes then, open up a cmd32.exe window and type:
java -version
(that same question was asked in comment #9)

A last thing might be that you have a 64-bit Octave version but your Java version is 32 bits.  But: I've never heard (or experienced myself) bad consequences of this, but maybe on your box things are different.

Some things that may play a role:

  • In step #1 you didn't load the io package, so little wonder that neither the testscript nor the binary modules can be found and you had to go to the io package directory to run the script; but the binary modules are still not in the path. ==> What happens if you do 'pkg load io' before running that testscript (w/o going to the package directory)?


  • When you try to do spreadsheet I/O the first time after starting Octave, the io package functions usually try to see if Java is present, using a private function that in turn calls 'javachk ("jvm")' (the same command I asked you to try above). If Octave crashes at that stage there's a problem with Octave versus Java and the io package can't do anything about it. However, I asked you to explicitly specify a spreadsheet I/O interface (i.e., the "OCT" interface) and in that case, for that interface, that Java check will be skipped entirely. So the io package doesn't use Java at all then.


  • We've had several issues with Octave and Java, there are quite a few reports on the bug tracker about it. What you can try is uninstalling Java and then install another Java release.




Philip Nienhuis <philipnienhuis>
Group Member
Wed 27 Mar 2024 05:20:35 PM UTC, comment #11: 

Sorry for the lack of info; I reinstalled 'octave-9.1.0-w64-installer.exe' and performed following test:

TEST #1- after renaming "C:\Program Files\GNU Octave\Octave-9.1.0\mingw64\share\octave\packages\io-2.6.4\__init_io__.m" to something else results:

濰ǂ
GNU Octave, version 9.1.0
Copyright (C) 1993-2024 The Octave Project Developers.
This is free software; see the source code for copying conditions.
There is ABSOLUTELY NO WARRANTY; not even for MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  For details, type 'warranty'.

Octave was configured for "x86_64-w64-mingw32".

Home page:            https://octave.org
Support resources:    https://octave.org/support
Improve Octave:       https://octave.org/get-involved

For changes from previous versions, type 'news'.

>>  pkg list io

Package Name  | Version | Installation directory
--------------+---------+-----------------------
          io  |   2.6.4 | C:\Program Files\GNU Octave\Octave-9.1.0\mingw64\share\octave\packages\io-2.6.4

>> io_testscript ("oct", "io-test.xlsx")


Testing interface OCT using file io-test.xlsx...

 1. Initialize arrays...
 2. Insert first empty sheet...
'col2num' undefined near line 49, column 10

ans =
{
  [1,1] = OCT
  [1,2] = -
  [1,3] = -
  [1,4] = -
  [1,5] = -
  [1,6] = -
  [1,7] = -
  [1,8] = -
  [1,9] = -
  [1,10] = -
  [1,11] = .xlsx

note that 'io_testscript ("oct", "io-test.xlsx")' must be run from 'C:/Program Files/GNU Octave/Octave-9.1.0/mingw64/share/octave/packages/io-2.6.4', otherwise it delivers:

>> io_testscript ("oct", "io-test.xlsx")

error: 'io_testscript' undefined near line 1, column 1

>>


TEST #2- after restoring "__init_io__.m" name:
same behaviour as test 1 before, except:

>> pkg load io -> causes Octave to instantly crash
>> pkg test io -> causes Octave to instantly crash


TEST #3- debugging function pkg load io:

>> pkg load io

stopped in methods at line 72 [C:\Program Files\GNU Octave\Octave-9.1.0\mingw64\share\octave\9.1.0\m\miscellaneous\methods.m]
72:         mtds_str = javaMethod ("getMethods", "org.octave.ClassHelper", obj);
debug>

the workspace is as follows:
# name:            # type:           # size     # value
havesigs           bool              1x1        0
mtds_list          cell              0x0        {}
obj                string            1x17       java.lang.Runtime
showsigs           bool              1x1        0
valid              bool              1x1        0

execution of line 72 causes Octave to instantly crash. javaMethod is compiled and I cannot debug further.






Giorgio B <boomer>
Thu 21 Mar 2024 01:40:27 PM UTC, comment #10: 


> I have the same problem on work PC ...

Exactly what problem, occurring when?

In the comments below I have asked to carry out some trial code. Did you do that, esp. from comment #3 and comment #5, and what results did you get?

As I wrote in comment #8, "work PC's" are often suspect because many IT depts. blacklist executables and .dlls. That's why I ask you specifically to try what I asked in comment #5.

Philip Nienhuis <philipnienhuis>
Group Member
Thu 21 Mar 2024 10:56:40 AM UTC, comment #9: 

What does `version -java` return on the affected PCs?

Markus Mützel <mmuetzel>
Group administrator
Thu 21 Mar 2024 10:40:48 AM UTC, comment #8: 

I have the same problem on work PC (win 11 os) with octave 8.3, 8.4 and 9.1, all 64bit, but the 8.3.0 32bit (the most recent I could find) runs smoothly. I hope this info is helpful

Anonymous
Wed 03 May 2023 09:07:12 PM UTC, comment #7: 

Coming back to this report, I didn't note you had commented further because I wasn't in the 'cc' list. Aplogies for that, added myself now.

The fact that 'col2num' (a binary module, actually a disguised .dll file) is undefined is telling. It can mean that:

  1. either the io package is wrongly installed
  2. or somehow col2num from a different and binary incompatible Octave release is in the path
  3. or (most probably) your IT dept. allows only explicitly 'whitelisted' SW to load and hasn't whitelisted the binary modules of the io package (and maybe other packages).

As to the latter, at my work they used 'Traps' SW to accomplish the latter. But there's more SW that can do the same.

Advice: talk to your IT dept. to find out if they block non-whitelisted SW.

Philip Nienhuis <philipnienhuis>
Group Member
Mon 10 Apr 2023 05:15:11 PM UTC, comment #6: 

So, as I said previously, when it crashes there is no error message.  The program just immediately closes.  I have screen recorded the crash on a fresh install, but was unable to upload it here (it should be within the file size limit).

With the _init_io_ file renamed, I am able to load like previously.  Calling io_testscript yields:

>> pkg load io
error: '__init_io__' undefined near line 1, column 1
error: called from
    C:\Program Files\GNU Octave\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4\PKG_ADD at line 1 column 1
    load_packages_and_dependencies at line 56 column 5
    load_packages at line 53 column 3
    pkg at line 639 column 7
>> io_testscript ("oct", "io-test.xlsx")

Testing interface OCT using file io-test.xlsx...

 1. Initialize arrays...
 2. Insert first empty sheet...
'col2num' undefined near line 49, column 10

ans =
{
  [1,1] = OCT
  [1,2] = -
  [1,3] = -
  [1,4] = -
  [1,5] = -
  [1,6] = -
  [1,7] = -
  [1,8] = -
  [1,9] = -
  [1,10] = -
  [1,11] = .xlsx
}

>>

I assume col2num is undefined because the io package does not load properly without the init function?

The next thing I tried is to remove all of the files and start adding them back one by one as errors were thrown for missing functions. As soon as '__chk_java_sprt__.m' was added back, I no longer got an error to debug, and instead the program crashed.

>> [numdat, txtdat, rawdat] = xlsread(filename);
error: 'xlsopen' undefined near line 379, column 13
error: called from
    xlsread at line 379 column 11
>> [numdat, txtdat, rawdat] = xlsread(filename);
error: 'getinterfaces' undefined near line 308, column 16
error: called from
    xlsopen at line 308 column 14
    xlsread at line 379 column 11
>> [numdat, txtdat, rawdat] = xlsread(filename);
error: '__chk_java_sprt__' undefined near line 109, column 21
error: called from
    getinterfaces at line 109 column 19
    xlsopen at line 308 column 14
    xlsread at line 379 column 11
>>

Again, I'd also like to reiterate that this issue has occurred on two separate work PCs (both with fresh installs done as well), but works perfectly fine on my personal PC.  Both work PCs have the same anti virus software installed by IT, where my personal PC does not.  Further, this package was working on those PCs until recently.  All PCs have up to date Windows, so the Anti-virus is the only thing I can think of that is common between work PCs and not personal PC.

Anonymous
Wed 05 Apr 2023 06:54:19 PM UTC, comment #5: 

Thanks, yes, the first error is to be expected, obviously.

Then the crashes.

'pkg test io' doesn't test the spreadsheet I/O at all so I'm suspicious that xlsread.m itself is not involved but that there are other errors provoking the crash. I'm especially suspecting that you have some io package functions, or maybe core Octave functions, in your load path shadowing io package, or core Octave functions.
Can you please copy verbatim the exact error messages you get into the next comment? Without those we can't do anything to help you further.
(use +verbatim+ and -verbatim- tags each on a separate line enclosing the copied output).

Next, xlsread crashing .... same thing, exactly what error messages do you get when using xlsread?

What spreadsheet I/O "interface" do you use?

The (default) 'OCT' interface doesn't need Java or other support software so I'd first want to know what happens when using that.
Hint:
You can try the following:

io_testscript ("oct", "io-test.xlsx")

Pls report back the exact error messages.

Philip Nienhuis <philipnienhuis>
Group Member
Tue 04 Apr 2023 07:38:26 PM UTC, comment #4: 


>> pkg load io

error: '__init_io__' undefined near line 1, column 1
error: called from
    C:\Program Files\GNU Octave\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4\PKG_ADD at line 1 column 1
    load_packages_and_dependencies at line 56 column 5
    load_packages at line 53 column 3
    pkg at line 639 column 7

>>
>> pkg list io

Package Name  | Version | Installation directory
--------------+---------+-----------------------
          io *|   2.6.4 | C:\Program Files\GNU Octave\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4

So it seems to load it (apart from the obvious error of the file being missing)

pkg test io still seems to cause a crash though.
xlsread also causes a crash (because it tries to access Java).



Anonymous
Tue 04 Apr 2023 05:22:00 PM UTC, comment #3: 

(Wild guess)  maybe related to the io package initialization process?  and/or the Java subsystem? (which is accessed during io package initialization)

In the command line, type:
cd (strrep (which ("oct2xls"), "oct2xls.m", ""))

and then (assuming you're running the GUI) in the file browser, right-click the file '__init_io__.m' and rename it to something else.
Next, exit and restart Octave, try again to do 'pkg load io', and please report back if you still see the crash.


Philip Nienhuis <philipnienhuis>
Group Member
Tue 04 Apr 2023 04:46:20 PM UTC, comment #2: 


>> pkg list io

Package Name  | Version | Installation directory
--------------+---------+-----------------------
          io  |   2.6.4 | C:\Program Files\GNU Octave\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4

>> pkg load io


And then Octave crashes.  No error to debug, program doesn't stop responding, just closes.

After further investigation, it may be firewall/anti-virus related.  This issue is happening on multiple work PCs, but not on my personal PC.

Anonymous
Wed 29 Mar 2023 05:57:44 AM UTC, comment #1: 

Loading the package is working for me:

>> pkg list io
Package Name  | Version | Installation directory
--------------+---------+-----------------------
          io  |   2.6.4 | C:\Program Files\GNU Octave\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4
>> pkg load io
>> pkg test io
Testing functions in package 'io':

Integrated test scripts:

  ..\packages\io-2.6.4\x86_64-w64-mingw32-api-v58\cell2csv.cc-tst  pass    2/2
  ..e\packages\io-2.6.4\x86_64-w64-mingw32-api-v58\col2num.cc-tst  pass   12/12
  ..\packages\io-2.6.4\x86_64-w64-mingw32-api-v58\csv2cell.cc-tst  pass   18/18
  ..e\packages\io-2.6.4\x86_64-w64-mingw32-api-v58\num2col.cc-tst  pass    3/3
  ...1.0\mingw64\share\octave\packages\io-2.6.4\calccelladdress.m  pass    9/9
  ..ctave-8.1.0\mingw64\share\octave\packages\io-2.6.4\fromJSON.m  pass  152/152
  ..ve-8.1.0\mingw64\share\octave\packages\io-2.6.4\object2json.m  pass    2/2
  ..-8.1.0\mingw64\share\octave\packages\io-2.6.4\read_namelist.m  pass    3/3
  ..\Octave-8.1.0\mingw64\share\octave\packages\io-2.6.4\toJSON.m  pass   56/56

Fixed test scripts:


Summary:

  PASS                              257
  FAIL                                0

See the file C:\Users\Markus\fntests.log for additional details.

34 (of 39) .m files have no tests.

Please help improve Octave by contributing tests for these files
(see the list in the file C:\Users\Markus\fntests.log).

>>


What does `pkg list io` return for you? Which exact commands are causing the crash for you?

Markus Mützel <mmuetzel>
Group administrator
Tue 28 Mar 2023 10:37:48 PM UTC, original submission:  

Trying to load the io package causes Octave to instantly crash on Windows.  This started happening recently. 
Trying to use the xlswrite function (without having the package installed) seemed to indicate that the crash seems to happen during _chk_java_sprt_ function.

Anonymous

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by dasergatskov (Posted a comment)
  • -email is unavailable- added by boomer (Posted a comment)
  • -email is unavailable- added by philipnienhuis
  • -email is unavailable- added by mmuetzel (Posted a comment)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only group members can vote.

     

    Follow 6 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2024-04-10 philipnienhuis StatusNeed Info Invalid / Not an Octave Bug
        Open/ClosedOpen Closed
    2024-03-29 philipnienhuis Summary[octave forge] (io) package crashes Octave Invoking Octave's Java subsystem crashes Octave
    2023-05-03 philipnienhuis Carbon-Copy- Added philipnienhuis
    2023-03-29 mmuetzel StatusNone Need Info
        Summaryio package crashes Octave [octave forge] (io) package crashes Octave

    Back to the top

    Powered by Savane 3.13-d3ae.
    Corresponding source code