Hello.
I am currently doing the final testing of the updated VFP Modeling Connection for VUML 2.9 that "hopefully" will be available tomorrow or Monday.
First up I'll provide y'all with a quick summary of the changes/enhancements.
Visual FoxPro Modeling Connection Enhancements
You now have the option when generating code to move either VUML attributes and operation descriptions or comments to VFP property and method descriptions. This selection is persistent for the model and stored in the Registry.
You now have the option when reverse engineering whether VFP property and method descriptions are moved to VUML attribute and operation comments or descriptions. This selection is persistent for the model and stored in the Registry.
You now have the option when reverse engineering and moving VFP property to VUML to clear the item (attribute and operation comments or descriptions) not selected.
You now have the option when reverse engineering and moving VFP user-defined method code to VUML operation code. This selection is persistent for the model and stored in the Registry.
Visual FoxPro Modeling Connection Corrections
Generating code arrays defined in the VUML with the format <
> is now authomatically converted to use VUML multiplicity, thereby allowing reverse engineering of the arrays. The system informs you of this change during the code generation process.
An issue causing array multiplicity not to be reverse engineered has been corrected.
An issue causing array visibility not to be reverse engineered has been corrected.
An issue causing property and method visibility not to be recognized during a second reverse engineering has been corrected.
An issue causing the display of an invalid property when generating code in the case where there are no properties defined in VUML has been corrected.
An issue causing the corruption of inheritance information during code generation has been corrected.
The main other issues that we have on our plate are.
1] OLEPublic not transported either way.
2] Method code is only transported to VUML from VFP not the opposite, and only for the class the method is created in.
3] Attribute/Property and Method/Operations are not deleted in either direction.
4] Support for .prg based classes.
5] Parameter information is not transported either direction.
Interestingly enough #3 is probably the easiest to implement, but to this point we have not done so by design to avoid breaking code. I will be doing an assesment of what is required to make all of these safe in the near future.
I do thank all of you for your input and support in this process.
Ok.
Bye.
>Rob,
>
>Actually, I am about --><-- that close to buying a copy. I run a smallish (right now) shop with about 5 developers and I'm looking to get a copy for myself personally to spend a few months learning, and then when we get a few more developers I plan to get out a whip and force everyone to learn UML. <g>
>
>Having said that (and scared off all the prospective employees here on the UT <g>) I dod have two technical questions I'd like to ask you if you don't mind.
>
>The first was brought up by David Frankenbach where apparently the VUML <--> VFP cycle isn't 100%. That is, if a change is made in VFP after making a change in VUML then that doesn't get propogated back I may not be correctly representing Dave here. There's a thread over on www.west-wind.com with the title, "Re-RevEng doesn't update model" you might want to take a look at and perhaps respond to.
>
>Secondly, yesterday we attended a Microsoft .NET rollout here in Salt Lake City and I had a chance to talk with the folks from Rational. I was curious to know if both models exported text. Apparently they both export XML. If so, do you have a sense on how hard it would be to create a diagram in VUML and import it to Rational. That is, unless you do everything they do as well. We are in a situation where we will be using both VFP and some .NET stuff and it would be nice to be able to maix & match.
>
>Thanks..
>
>
>>Yes, you and others were correct in admonishing me regarding the inappropriate remark that I made. It was not intended to be a threat, veiled or otherwise. My grievance was with Christoph (who I still contend made several untrue accusations - hacking into his computer, blocking his IP) in public, not with the VFP community. I am not really advocating that everyone go to the competition, though that is obviously an option that all of us living in a free society have if we do not care for a product or its vendor.
>>I have also considered the remarks that a couple of you made about there perhaps being something bigger than the initial problem that stimulated the reaction. The only thing that I can come up with (aside from being tired from working too much) is some level of frustration with the apparent difficulty in trying to sell a UML for VFP tool into the VFP market. We at Visual Object Modelers, and Markus and his people at EPS Software, have worked long and hard to try and provide a good UML tool for VFP development but it has proven to very difficult and not as successful as originally hoped for. I am not aware of any competitor that provides VFP support in a UML tool and have been wondering why not. Perhaps they have determined that the market is not big enough.
>>
>>Rob
