Hi all,

I'm running into strang ROPS behaviour when using the latest compiler in svn. Is there a compiler switch required for it to work ok? Using revision 3823 I have no problems.

The parameters in the script when passed to actual FPC code is not working well.. this was an old bug... parameter order or I don't know what...

Anyone has any clue for me?

That are 1300+ revisions. Try doing a binary search. That will take 11 tests.

Peter

2006/10/31, Peter Vreman :

How do a do such a thing?

Update to rev. 3823+650 e.g. something like 4500 (svn up -r 4500). Try with this one, if it fails try with the rev. in the middle of 3823 and
4500. If it works with 4500, try the rev. in the middle of 4500 and HEAD. And so on ... with 11 of such tests you find the revision causing you trouble.

Alexandre Leclerc wrote

> That are 1300+ revisions. Try doing a binary search. That will take 11 > tests. How do a do such a thing?

You know binary search right ? Use svn diff -r X, with X varying between working and newest revision, until you found X-1 working, X not.

Micha

Use the binrary search algorithm to find the the revision that break it. Everytime update svn with "svn up -r <rev>" to update to the revision to test.

Peter

Have you tried ROPS svn revision 19?

Jesus Reyes A.

___________________________________________________________ Do You Yahoo!?
La mejor conexión a Internet y <b >2GB</b> extra a tu correo por $100 al mes. http://net.yahoo.com.mx

2006/10/31, Peter Vreman :

Ok, ok, I know what you guys mean now... Yeah, this is the 'QuickSort' algo (how I use to call it) and I'm using this technique too. I did not know what was 'binary search'. :)

Best regards.

2006/10/31, Jesus Reyes :

Yes and it solved my 'Self' problem (as you can read on rops forum). But when I compile with latest compiler my script is not behaving completely right... Just with a different compiler version. :(

The passing function prarameters are not good... this was an issue couple months ago that Carlo fixed. But now it comes back. Maybe this was a fix on a FPC bug? or an FPC bug has been introduced...

The issue are the parameters when calling a Delphi function from the script.

2006/10/31, Alexandre Leclerc :

For now, the only thing I'm sure about is that with version 2.0.4 of the compiler it is working correctly with ROPS for this problem.

For now, the only thing I'm sure about is that with version 2.0.4 of the compiler it is working correctly with ROPS for this problem.

Can't you try the binary search to find the revision causing trouble? Are there any fpc defines in the ROPS sources? Where are the ROPS sources available?

2006/11/1, Florian Klaempfl :

Binary search is complex because I have difficulties to get a FPC revision that will be able to compile latest Lazarus version... (unless I do not need to recompile my Laz each time but simple replacing the fpcsrc & pp folders?)

ROPS is available from subversion:
svn://svn.remobjects.com/pascalscript

Any-way I'll try again binary search, but this morning I wasted 3 hours for no results... and this is really slow. Some FPC revisions were not compiling, other were not able to compile lazarus... :S

What do you expect from us? That we spend hours and hours debugging your problem?

Peter

Alexandre Leclerc wrote

Binary search is complex because I have difficulties to get a FPC revision that will be able to compile latest Lazarus version... (unless I do not need to recompile my Laz each time but simple replacing the fpcsrc & pp folders?)

Does ROPS depend on Lazarus? If not, there's no reason to recompile your Lazarus every time.

Jonas

2006/11/1, Peter Vreman :

No Peter, I expect no specific help from you if you have no help to give. I did say I was going to try again, having more light on some issues so that I can more efficiently find the problem. But if you absolutely want to do it, well, go for it! :)

Also, it might not be totally my problem. I'm using FPC & ROPS and something somewhere is not working. The only thing I know is that: 2.0.4 + ROPS = ok;
r3823 + ROPS = ok;
HEAD + ROPS = not ok.

Therefore my first intuition is that something changed in FPC. This is the origin of this thread in case someone whould have known... Now I'm trying to find the revision... painfully. But with the comment of Jonas, I'll get it faster.

Regards.

2006/11/1, Jonas Maebe :

Only for IDE integration, but I think I can skip that part. I'll save time now doing the binary search. Thanks.