Jump to content

Jason D.

  • Content Count

  • Joined

  • Last visited

Everything posted by Jason D.

  1. Is there a way to programatically verify the versions of the installed VIPM packages? We want to test that our build machine contains the approved version of each installed package before building our labview apps. It looks like we can parse the ogp fiile names in C:\ProgramData\JKI\VIPM\databases\LV 11.0\, but is that the correct way?
  2. I installed VI Tester in LabVIEW 2011 and there is no toolbar for it in the Project Window. Is that a bug, or an incompatibility or is there just something wrong with my install? Thanks Jason
  3. You're right, the check box is there. However I still think this is a minor bug in VIPM. If it could just wait the requested timeout then there would be no need for a workaround.
  4. I don't think the popup can be suppressed. But if VIPM could just wait longer to connect, everything would work. Even though there is a 90second timeout, VIPM puts up an error message box before I can even hit the "launch" button on the NI popup. (it's not really urgent; i'm just suggesting an improvement for a future release).
  5. Not the LV 8.6 beta, but the LabVIEW Software Engineering Tools beta (Trace Execution Toolkit and Unit Test Framework).
  6. I'm trying to transition to LabVIEW 8.6, and I just installed VIPM 2.0.2 as well. I'm trying to add LV 8.6 to VIPM, and I'm in the VIPM - Configure LabVIEW Port screen. I press the test button, and it fires up LabVIEW, which brings up an "evaluation software" dialog box, with the number of days remaining on the beta licenses. Before I can even click on "Launch LabVIEW", VIPM gives a failure message saying it can't connect to LabVIEW 8.6. Then I tried to increase the timeout from 90sec to 600sec. It still failed right away. I didn't prove that the beta tookits are the problem, but it worked fine on my other (albeit non-Vista) machines. After this, LabVIEW had actually started up. When I tried to hit TEST again, then it said LabVIEW was already running. I hit the "Quit LabVIEW" button and VIPM became unresponsive. I waited a minute or five and then killed it with the Task Manager. I was able to make the connection by manually fixing the LabVIEW VI Server configuration. Also, the error message says "Make sure you are allowing access to VIPM by specifiying "*" in the allowed list. It would be a lot safer to tell people to allow "localhost" rather than "*".
  7. Hi Jim: I tried to use XML to persist arbitrary data types to disk. However, the EasyXML file VIs convert the random EOL characters found in my flattened strings so the data read from the file is not identical to the data saved. I think this could be classified as a bug. My original enum was slightly larger (only slightly) and caused LabVIEW to crash every time. The attached example merely throws error 74. The easiest workaround is probably to read and write the files with separate file I/O calls which call out binary mode instead of text mode. Have you given any thought to a way to implement XML CDATA? That might be a better way to handle the storage of flattened LabVIEW data. If you have any other good ways of storing arbitrary LabVIEW data without needing to wire in the data type, I would be interested. I hope you are enjoying NI Week! Jason Enum_Conversion.vi
  8. Jason D.

    Error 1

    Here are the differences. I am using LV 8.5 ogrsc_dynamicpalette-0.17-1 jki_lib_easyxml-1.0-1 VIPM says I am up to date, so I don't know how to get the newer version of EasyXML you have. I have not yet backed out my OpenG fixes, but I will try that just to be sure.
  9. Jason D.

    Error 1

    Hi Jim: I'm glad to see it is working for you. I just get Error 1 and an empty array. I will check all of my packages later., but I haven't updated in a few weeks. I have put in my own fixes for this OpenG issue, so I will look closely at that, and maybe try with a clean version of OpenG. [Thanks for removing my default data from your site!] Jason
  10. Jason D.

    Error 1

    OK, I have a much more tractable example which reproduces the problem . Any ideas? Two_way_XML_List_Conversion_Fails.vi
  11. Jason D.

    Error 1

    EasyXML threw an Error 1: That's all well and good, but my data structure is huge, and the XML input string was created by a few previous calls to EasyXML with a very similar data structure. If EasyXML could add run-time information with the label of the offending control, then I could get this debugged a lot faster. In the meantime, do you have any other advice? (I can't post the code here, sorry).
  12. [Note from Moderator: This post was split off of this topic] Hi Jim: I added an item to your examples for a list of scalars, and I think it would help others to see the example. I was really happy it worked! Jason Scalar_Element_with_Attributes_Example.vi
  13. The default formatting of %#g only provides NI's legacy default of six digits of precision. That means that most doubles are pretty much cast to singles when they go through EasyXML. Also, digits of precision is inferior to significant digits, because with the former, the number of digits preserved is dependent on the magnitude of the number. For example: 100.12345678 => 100.123457 and with 1.0012345678 => 1.001235 It's floating point for 's sake. The whole idea is that magnitude doesn't affect precision. Is there any way you can put %^#_14g into the numeric formatting? You could also use %^#_7g for singles, but that's icing on the cake. The NI Help recommends using only 13 and 6 digits, because they don't want you to use the last uncertain digit. However, the application (including EasyXML) should preserve that digit so that precision is maintained as long as possible throughout computations. I also like the caret (^=engineering notation) because it makes the exponential numbers slightly more readable, and some measure of human readability is part of what XML is about.
  14. The instructions for Vista installation did not work for me. I got the error 740 three times. However it is very easy to right-click on the installer EXE and choose "Run As Administrator". That worked just fine.
  15. If you try to uninstall a package, the default is "include dependencies" and all the related packages are checked. I assume this would break many of the other packages one might be keeping. I would suggest keeping the dependencies unchecked by default for uninstall operations.
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.