Welcome to our forums
If you are using the Visual Studio plugin and you have a bug or issue can you post a description in this thread. If you are using the plugin and have no problems can you also post in this thread so we can get a feeling for how many people are using it.
I have a C++ program that calls my Silverfrost DLL. Is there a way I can debug the Fortran using Visual studio 2015 (as I cannot use sdbg64 in this scenario)? So basically I start debugging the C++ program in Visual studio and when it gets into a call in the Fortran DLL, it will continue debugging (Fortran inside VS). Is that possible?
I don't know but if you build your DLL using /DEBUG on the command line then you could try stepping into a Fortran routine when stepping through your C++ code.
I have expressed my point of view already in earlier threads, where I have also given some remarks concerning the present Visual Studio plug-in. Just to summarize: Making Fortran a .NET language would certainly be one promising route of using (old) Fortran programs in a modern environment. A revised and consolidated ftn95 plug-in for the latest Visual Studio version would be ideal for further testing this route. Visual Studio is an excellent IDE and offers an easy-to-use GUI. Klaus
Paul, Is this resolved in the latest FTN95 8.40? .... The long awaited support fix?..
Can someone who is a Visual Studio enthusiast or user explain the benefits of the system relative to the other options in FTN95 please?
Given that FTN95 has its own IDE (Plato) which is fully integrated with the compiler and its help file system, why is VS’s IDE any better? (Klaus says: “Visual Studio is an excellent IDE and offers an easy-to-use GUI.” Doesn’t Plato? Are we talking about the user interface for the IDE or the ability to build a program with a GUI?)
If it is a case of building a program with a GUI, then given that FTN95 has Clearwin+, does the Visual Studio and/or .NET (if they are different) approach give the programmer a quicker route? Or a better finished result? e.g. with VS, can you do things that there are no equivalents for in Clearwin+, such as movable toolbars or a “ribbon”?
Klaus also says that VS is a “promising route of using (old) Fortran programs in a modern environment.” OK, it’s an approach – one of several. But is it better (or more “promising”) than any other routes (console, or via Winapp/Clearwin+) outside VS? For example, is there a 64-bit VS version? Isn’t the compiler effectively the same, with the same error messages, or is it better when integrated with VS? If so, in what way? Might it be faster, or generate smaller code? Or not?
In UK currency, VS 2017 Enterprise is GBP3k, Pro is GBP400, and Community is free. What don’t you get in the free version?
Finally, is VS 2017 better in any significant ways than previous versions that FTN95 will integrate with? If so, what are they?
Eddie
Edit: 1st January 2019. Given that there is now a Visual Studio 2019, perhaps there is a question over Visual Studio 2017 support - is there?
Excellent contribution, John, but nary an answer to any of my questions. Of course, that's because no-one is working across the weekend and into the New Year, or maybe ...
John,
Your latest post started so well ! 'with refusal of companies usually to even consider running 2 software of 'same type''
Then followed it up with 'I mean what does anyone need more than Office 2003'
The company I previously worked for refused to allow me to USE Excel 2007, which was the version that allowed substantially larger sheet sizes. 'We have standardised on Office 2003' was their response, with no idea of why I might need a larger sheet size.
My response was I bought and installed Excel 2007, as I had administrator rights for FTN95 installs, which they also did not understand, but I had been using since before they became an IT department.
As an engineer, an IT department is not my friend !
John
John S,
You said 'you shouldn't be using spreadsheets so big' I use them with pivot tables; the best way to analyse results.
To both of you, thanks for your contributions, but the questions were really very serious and if there were appropriate answers to many of them might well consider using the visual studio approach. I suspect, however, that the answers are that it’s somewhat different, with one or two benefits but actually rather a lot of disbenefits, and so it is worth persevering with what I’m using already.
Remember please that this is a VS thread (said by someone who regularly goes off-topic) and I’m interested in answers to the VS queries. (I’m also interested in this discussion, but it ought to be in ‘General’).
Eddie
If you are using .NET and FTN95 then the VS IDE becomes very useful (if not indispensable) for debugging.
VS is better than Plato in certain peripheral areas (for example when it comes to editing images). if there other areas directly related to FTN95 where VS is considered to be better then please let me know.
On a serious note, it's over a week, and no-one has offered us the slightest reason to use VS/.NET, except Paul, who points out that the VS IDE is good for debugging .NET FTN95 programs. That is something of a given, I should think.
Well, I suppose it would be, if FTN95 would install in the latest VS versions!
Eddie
Quoted from LitusSaxonicum ... the VS IDE is good for debugging .NET FTN95 programs. That is something of a given, I should think.
Well, I suppose it would be, if FTN95 would install in the latest VS versions
I am afraid that keeping up with the frequent VS updates that MS releases can consume a lot of attention and effort. Even mighty Intel had a major problem for a number of weeks between August and October, see https://software.intel.com/en-us/forums/intel-visual-fortran-compiler-for-windows/topic/784330 .
mecej4,
I suspect that the difficulties of continuing to support VS / .NET integration with FTN95 make it a complete and utter waste of time, but I would like to see the answers to my questions, as I could be convinced to change my view in the right circumstnces.
It is interesting that Intel found it difficult to respond to changes in VS. I think we shouldn't expect FTN95 integration any time soon.
I don't think I will lose much sleep over it.
Eddie
It is worth noting that the various versions of Visual Studio generally provide incremental updates to the IDE and from my perspective the changes can sometimes be simply cosmetic and unwelcome.
I use VS for C/C++ programming (the language used in the construction of FTN95, ClearWin+, Plato etc.) and I have recently started using the 2013 version. Whereas 2013 has some new features (for C++) that I really like, it also has some automatic responses that I find really irritating. The result is that on balance there is little advantage (for me) to move up to 2013. I haven't even contemplated upgrading to 2017.
For FTN95 programmers, one advantage of using Plato is that Silverfrost has access to the whole of the code which means that it is relatively easy to fix bugs and to respond to requests for improvement.
Approximately 4 weeks ago, Robert had sent me the alpha version of the new Visual Studio (VS) 2017 Fortran plugin for testing. This version works very well, and -for the moment- only minor issues need further attention. For my work, this is an essential contribution for using Fortran within VS and I would like to acknowledge Robert’s work. Just a comment to the question why using VS: For large Fortran projects I prefer VS instead of Plato. Advantages are
Klaus, While not explicitly answering my questions, your recent post clarifies some issues for me in a helpful way. If I may paraphrase your answers:
... continued (for some reason the Forum posted my incomplete message and won't let me edit it).
So (1) is a matter of taste, fair enough. As far as (2) goes, what is the definition of 'big', and does PLATO actually have a size limitation? (3) Seems fair enough, but then that doesn't apply to anyone who uses just Fortran. I can see that (4) could be a real issue for some. Many thanks for the insight.
Eddie
Early versions of Plato had problems with dependency checking. This was particularly apparent for large projects. The current Plato works well with large projects as far as I know.
Last summer, I had some problems with Plato to define a Fortran project consisting of approximately 550 subroutines or functions. I had to include the files in several 'packages' of about 100 routines. After I had included such a 'package' the dependency checker started to work automatically, which I did not noticed immediately. Any attempt to include further files ended in a disaster. So it took some time to build up the project.
I also had a problem with Plato's dependency checker when I modified a global variable in a module, which is used in more or less all routines.
I am glad to hear that these problems have been solved and I will have a look to Plato again when the latest ftn95 PE version will be released.
I completely agree with your comments, Eddie. Of course, we should always keep things as simple as possible.
Klaus