forum.vdsworld.com

FreezingFire · **Admin Team** Joined: 23 Jun 2002 Posts: 3508

Thanks for keeping us all updated, Marty. Smile

jules · Posted: Fri Jun 27, 2003 6:02 pm Post subject:

'm not speaking for Commercial Research on this, and I know that Emmanuel has devoted a lot of his valuable time on this issue, but I take exception to the statement that there is a "problem with VDS 5".

As the person who did most of the development work on VDS 5 I can state that nothing was done to knowingly break any existing VDS extension. All of the extensions I personally have written, and others that I have purchased and use, worked perfectly under VDS 5 without requiring any modification and have done from the outset. The extensions interface has not changed, and works in the same way as it did previously.

However, VDS 5 has a brand new IDE and works in a different way (for example, VDS50RUN.DLL now provides script execution for it whereas previously the IDE had a compiled-in copy of the script engine) which could mean that errors in extensions such as memory leaks or unfreed resources that went undetected before are now producing problems.

The bottom line, in my opinion, is that there is actually no requirement for VDS 5 to support existing extensions. Each time a new version of Delphi comes out new versions are needed of all the components. A lot of things would have been possible if it had been possible to change the extensions interface in a way that broke all existing extensions, to allow new features. So if a few extensions don't work with VDS 5, I think it's up to their developers to update them. Instead of criticizing VDS 5 for not working with a few extensions, you should be glad that it still does work with most of them.

Just my opinion, not Commercial Research's or anyone else's...

Hortalonus · Posted: Fri Jun 27, 2003 6:19 pm Post subject:

marty · Professional Member Joined: 10 May 2001 Posts: 789

Ok that maybe true and thanks Jules, but what do I have to change to my extensions to make them work? Somebody to guide us VDS extensions developpers?

I do need the info from Commercial Research for that right?

Thanks

Skit3000 · Posted: Fri Jun 27, 2003 7:12 pm Post subject:

Marty, maybe you should download the Extension documentation and look if there are some changes in it in comparison with any older version. If so, that can lead you to your answer (I hope Embarassed

)...

marty · Professional Member Joined: 10 May 2001 Posts: 789

Thanks skit, didnt know that SDK was still valid, will look into that. Its dated 27 january 1997...

Tommy, PK or MindPower any comments about Jules post?

Since you guys are much more experienced in VDS extensions than me, Im a bit lost here. Dont even know were to start....

Thanks

FreezingFire · **Admin Team** Joined: 23 Jun 2002 Posts: 3508

If we ever do get what is changed it should be posted on the developer
forum too. Smile

http://developer.vdsworld.com/

PGWARE · **Web Host** Joined: 29 Dec 2001 Posts: 1566

I spoke to Mindpower last week about this and he informed me it looked like something within the INIT section was changed within VDS 5. He's fixed that within his code (he uses C) and he noted his dll's were working. I did the exact same thing in Delphi and it doesn't seem to work for me, while the dll's work on my computer they dont on others (1 person), but the dll has worked on other peoples computer without a problem -vdsobj02. The dll performs perfect with VDS 2 - 4 though on my system as well as others.

I can understand if there are memory leaks but if there are memory leaks VDS shouldn't be returning weird errors such as Arithmetic error when my dll does no math at all anywhere in it, and it certainly never returns that error message number within the dll back to VDS.

Also I think Nathan pointed out that the vdsxfn? (dll which returns a string reversed) also throws an error in vds5. Such an easy and small dll also is causing memory leaks? Maybe but I'd bet against it.

Last time I spoke with the people at Commercial Research I was told that the problems seem to occur only with dlls that place controls on the vds dialog, however Nathan's excellent work checking almost all dll's has proven this to not be the case. From what I've seen everything everyone (including myself) has said is just speculation as to what the problem is. I'm not rulling out that my dll's are at fault but I'm certainly not rulling out the problem is within VDS 5 either.

FreezingFire · **Admin Team** Joined: 23 Jun 2002 Posts: 3508

Here's what Johnny told me yesterday:

Tommy · Posted: Sat Jun 28, 2003 9:38 pm Post subject:

Although I haven't found a solution to the problem yet, I find that for my DLLs, the
errorcode is the maxpar value returned by the DLL, increased by 1.

For example with vdstwain.dll, maxpar is 100, so that it returns error 101 on the first
command call. If I change the value to 50, restart VDS 5 and retry, the returned error is 51.

jules · Posted: Mon Jun 30, 2003 8:18 am Post subject:

VDS 5 takes no notice of either the maxpar or bufsize variables set in the Init procedure. I seem to recall finding that VDS 4 didn't take any notice of one of them either. I can't see why the errorcode should be affected by this since it returns whatever value you set it to.

FreezingFire · **Admin Team** Joined: 23 Jun 2002 Posts: 3508

Just a thought: Why don't you compile a list of the things changed in
DLL processing since the release of VDS 4. This would really help people
who are lost. Smile

PGWARE · **Web Host** Joined: 29 Dec 2001 Posts: 1566

I noticed too that the maxpar variable + 1 is the returned error code. After Tommy pointed this out to me and I tried it sure enough that is the error number returned each time.

jules · Posted: Mon Jun 30, 2003 6:29 pm Post subject:

marty · Professional Member Joined: 10 May 2001 Posts: 789

Tommy has found interesting stuff about all these problems with VDS extensions + VDS 5. It may be a solution also! Great work Tommy!

http://forum.vdsworld.com/viewtopic.php?t=1678