Re: cimmof vs cimmofl
Quan Wang <[email protected]> Fri, 24 Aug 2012 16:45:29 -0500
| Newsgroups | gmane.network.open-pegasus.general |
|---|---|
| Message-ID | <OF6B985F9C.244F67F9-ON86257A64.007779E2-86257A64.0077815C@us.ibm.com> |
Much clear. Thanks! ============================================= Quan Wang HMC Development, IBM Austin Phone: 512-286-9146, T/L 363-9146 Fax: 512-973-4284 ============================================= From: Karl Schopmeyer <[email protected]> To: Quan Wang/Austin/IBM@IBMUS, Devchandra L Meetei <[email protected]> Cc: "Hwang, Johnny" <[email protected]>, Kirk Augustin <[email protected]>, "[email protected]" <[email protected]> Date: 08/24/2012 03:53 PM Subject: Re: cimmof vs cimmofl There are a couple of reasons for cimmofl the direct to repository compiler rather than cimmof the mof compiler that goes through the cimserver client interface and requires a running server. Note, however, that they are both the same compiler at heart and the only difference is the destination for the compiled objects. 1. Need a way to put some basic things into the repository when starting. Thus, loading qualifiers and the core classes is a straight forward task and does not require any interaction with the server itself, just the repository 2. It is logical to load a basic repository directly in the case of creating a new repository. Besides, early in the project we did not have enough of a running server that we could have used cimmofl and we needed a reposiitory to test the server components as they were built. However, using cimmofl bypasses the providers and directly installs anything defined into the repository. Since creating instances often goes through providers even before they are installed into the repository and providers may provide extra quality checks and may even do more than just creating instances in the repository (ex. creating filters, etc. for indications) we long ago adopted the policy that except for repository startup, users should go through cimmof. Further as we move into the future with a stable platform more and more users will want everything to be security filtered and only cimmof passes through the security components. Thus, except for basic repository creation we strongly encourage users to use only cimmof. Note that once a basic repository was created one could seriously consider disabling cimmofl. Calling cimmofl deprecated was probably too heavy handed but we were looking for ways to discourage people from using cimmofl except for the limited startup and that was what we adopted early in the project. We may look at rewording or reorganizing things a bit in the future to make some of these differences clearer to the user. Karl Thanks for your message at 03:15 PM 8/24/2012. Your message was: My understanding is, for the new installation and without CIM server start running, using cimmofl to construct the repository. cimmof is for compiling the cim classes with the CIM server running (of course it already has basic repository constructed. I just what to know if my understanding is correct. ============================================= Quan Wang HMC Development, IBM ============================================= From: Devchandra L Meetei <[email protected]> To: Kirk Augustin <[email protected]> Cc: "Hwang, Johnny" <[email protected]>, "[email protected]" <[email protected]> Date: 08/24/2012 02:39 PM Subject: Re: cimmof vs cimmofl I agree with Martin. though cimmofl is deprecated, it is used in make repository because, cimom need some of the classes to be present in repo to start to work properly. Regards --Dev On Sat, Aug 25, 2012 at 12:59 AM, Kirk Augustin < [email protected]> wrote: It has been awhile, but I believe the idea was the cimmofl is better if you are starting from scratch or have deleted the entire repository. But cimmof is what you should use if you are simply adding to or modifying the existing repository. Kirk Augustin 11821 NW McNamee Rd Portland, OR 97231 HM: 503-289-4356 From: "Hwang, Johnny" < [email protected]> To: " [email protected]" < [email protected]> Sent: Friday, August 24, 2012 11:52 AM Subject: cimmof vs cimmofl I’m using Pegasus 2.11.1 and Pegasus 2.10.0 . I notice in the pegasus/mak/configschema.mak file that CIMMOFCLI = $(PEGASUS_HOME)/bin/cimmofl However, cimmofl displays the message Warning: Use of cimmofl can corrupt the CIM Server Repository. cimmofl should only be used under very controlled situations. cimmof is the recommended OpenPegasus MOF compiler. Should cimmofl be changed to cimmof instead? Is there any reason why Pegasus is still using cimmofl during its make repository phase? Also, on Windows, I’m testing a new environment setup: Visual Studio 2010 and 2008 installed on the same machine. I compile with the VS 2008 version of vcbat. Executing the newly-compiled cimmofl then results in a buffer overflow. I see the same behavior with 2005 and 2008 installed, and using the VS 2005 version of vcbat. When there was only one version of Visual Studio installed, I do not have any problems. If you need more information, please let me know and I will flesh out the context more. Thank you all, and hope this helps, Johnny Hwang Office: 408.822.3231 Mobile: 510.304.7404 Karl Schopmeyer Inova Development Inc. 305 Spring Creek Village, Suite 475 - Dallas TX, 75248 USA EMAIL: [email protected] FAX: 1-972-239-0326 Phone 1-972-814-5581 Skype: kschopmeyer Skype Phone: (214) 556-5971