Sometimes you just cannot put it down!
Cliff

Mailing lists / lazarus / 1999-07
19 messages · 5 people ·
Reply by email Subscribe to this list Your subscription
Read-only archive. Use j and k to move between messages.
Sometimes you just cannot put it down!
Cliff
Shane I made a test program using Dispatch. Don't try and use it yet it appears to have problems. If I only define one procedure with a message it works fine. If I try to define a second one the compiler panics with a 9999 error.
For the FPC guys here is what blows. This gives and internal error 9999 and indicates it occurs at the 'end;' statement of the class definition. Now if comment out the second message and anything to do with DoClick2 the program compiles and works correctly.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++===
program msgtest;
{$mode delphi}
uses sysutils;
type
TMyObject = Class (Tobject)
protected
procedure DoClick(var msg); message 3;
procedure DoClick2(var msg); message 4;
public
procedure Test;
end;
procedure TMyObject.DoClick(var msg);
begin
writeln ('Called DoClick - msg = ' + inttostr(longint(msg)));
end;
procedure TMyObject.DoClick2(var msg);
begin
writeln ('Called DoClick2 = msg = ' + inttostr(longint(msg)));
end;
procedure TMyObject.Test;
var
msg : integer;
begin
msg := 3;
Dispatch(msg);
msg := 4;
Dispatch(msg);
end;
var
MyObject : TMyObject;
begin
MyObject:=TMyObject.Create;
MyObject.Test;
MyObject.Free;
end.Or use a compiler dated after 1999/07/06, i fixed it on the 5th. Peter
Peter Vreman wroteOr use a compiler dated after 1999/07/06, i fixed it on the 5th.
Okey dokey. I was planning on cycling tonight I report if it is still doing it. :-)
What bug is fixed? The Dispatch bug or the StrPCopy bug? Shane Fixed, it was a wellknown…
the crash in the compiler with multiple messages, see bottom of the message.
For the other 2 bugs i didn't see any source code
Peter
2 thoughts. 1) I did try to make code that didn't work, but it DID work.
2) I also get that message about LMessages can't be static linked. I don't get that on the box at work, but I do get sure why.
About making it an INC file. I guess I was going the same route as Delphi where it uses the Messages file to store stuff. I was planning on putting more into that file once I determined that it was appropriate to place it there. If we make it a .inc, I have no problem with that. But now you know why I did it this way. :-)
BTW, I did a new MAKE CYCLE from CVS but didn't notice any changes, though I didn't expect to.
Shane
>> "Michael A. Hess" 07/08 8:52 AM >>>
Peter Vreman wroteFor the other 2 bugs i didn't see any source code
Shane I didn't have time to put any sample together for Peter. Can you try and make a snippet that doesn't work? I won't get a chance until sometime late today or this evening. Besides I want to cycle a new compiler before I do anything else. :-)
Another thing I have problems with. When I link Lazarus I get an error message that it can't link LMessage statically. I'm not sure why it is doing this but since this is just a file of const declarations can't we make this an .inc file and then just use {$I lmessages.inc} in the controls.pp unit. Since controls is used by just about everything it will do have the same result. This might be a good idea for vclglobals as well. Personally I'd rather use globals then have a cluttered uses line in every file.
What bug is fixed? The Dispatch bug or the StrPCopy bug?
Shane
>> Peter Vreman 07/08 4:59 AM >>>
Fixed, it was a wellknown copy&paste bug :)
Peter
_________________________________________________________________ To unsubscribe: mail with
"unsubscribe" as the Subject
archives at http://www.miraclec.com/list_archives/lazarus
Interesting. I created a test app (which we all hate to do) to show Peter the problem, but it worked. That sucks.
The test app was:
Program Test;
{$mode objfpc}
uses sysutils;
Var
NewPChar : PChar;
StringVar : String;
Begin
Stringvar := '<hello>';
NewPChar := StrAlloc(Length(StringVar)+1);
StrPCopy(NewPChar, StringVar);
Writeln('StringVar = '+StringVar);
Writeln('NewPChar = '+NewPChar);
end.yet my source in gtkint.pp (procedure SetCallBack) does not work. Anyone have a clue?
Shane
>> Peter Vreman 07/08 8:42 AM >>>
the crash in the compiler with multiple messages, see bottom of the message.
For the other 2 bugs i didn't see any source code
Peter
_________________________________________________________________ To unsubscribe: mail with
"unsubscribe" as the Subject
archives at http://www.miraclec.com/list_archives/lazarus
Okey dokey. I was planning on cycling tonight I report if it is still doing it. :-) ====…
> Or use a compiler dated after 1999/07/06, i fixed it on the 5th. Okey dokey. I was planning on cycling tonight I report if it is still doing it. :-)
When cycling you should start with the 0.99.12a version, that should always work.
Peter
Interesting. I created a test app (which we all hate to do) to show Peter the problem,…
Shane Miller wroteInteresting. I created a test app (which we all hate to do) to show Peter the problem, but it worked. That sucks.
I have had that happen more times then I want to think about. Maybe I can try and debug it and trace down where the screw up is happening when I get back to my office this evening.
I have had my EZcgi code go bust with the latest compilers and I'm not sure why. I use StrPCopy alot so that might by why. :-\
<snip>
yet my source in gtkint.pp (procedure SetCallBack) does not work. Anyone have a clue?
Try making the test program just a little more complicated. The one thing that we are doing that this one isn't is running it inside a class method. That way you start bring objects and stuff into the mix.
the crash in the compiler with multiple messages, see bottom of the message. For the…
Peter Vreman wroteFor the other 2 bugs i didn't see any source code
Shane I didn't have time to put any sample together for Peter. Can you try and make a snippet that doesn't work? I won't get a chance until sometime late today or this evening. Besides I want to cycle a new compiler before I do anything else. :-)
Another thing I have problems with. When I link Lazarus I get an error message that it can't link LMessage statically. I'm not sure why it is doing this but since this is just a file of const declarations can't we make this an .inc file and then just use {$I lmessages.inc} in the controls.pp unit. Since controls is used by just about everything it will do have the same result. This might be a good idea for vclglobals as well. Personally I'd rather use globals then have a cluttered uses line in every file.
If you were planning on putting some more routines in there well then just leave it as it…
Or use a compiler dated after 1999/07/06, i fixed it on the 5th.
Peter
Interesting. I created a test app (which we all hate to do) to show Peter the problem,…
Interesting. I created a test app (which we all hate to do) to show Peter the problem, but it worked. That sucks.
Else you can also sent the need lazarus sourcecode and where it goes wrong in that code. (sent the code in private email)
Peter
When cycling you should start with the 0.99.12a version, that should always work. Peter
Peter Vreman wroteWhen cycling you should start with the 0.99.12a version, that should always work.
Oh I do. I have been it is just that I haven't built a new compiler since the morning of the 5th, before you fix. :-)
I'm just a few days behind. :-)
2 thoughts. 1) I did try to make code that didn't work, but it DID work. 2) I also get…
If you were planning on putting some more routines in there well then just leave it as it is. Maybe you could put a dummy routine in and see if the static link error goes away.
Shane I made a test program using Dispatch. Don't try and use it yet it appears to have…
Fixed, it was a wellknown copy&paste bug :)
Peter
2 thoughts. 1) I did try to make code that didn't work, but it DID work. 2) I also get…
Hello,
I've been thinking on what a Borland's Delphi for Linux will mean and I've changed my mind about the best way of creating a RAD for FPC on Linux, now that future Borland's product will be the one to beat.
Of course from my personal guesses on which will be the Borland's way. OTOH I've read a message by Michael Van Canneyt in FPC-Devel, about discussions in this list and what discussions made to Megido time ago.
Maybe he's right, I don't know. But I think that I better work alone on my very personal project until a more than reasonable amount of code is available to be published.
Since I don't intend to use gtk anymore, I won't be able to help Lazarus in any way, so it makes no sense to stay messing around here. I'm happy to see that my idea about gtk_fixed seems to have been finally useful.
I hope that I haven't been too annoying .
I wish you succeed with Lazarus... you surely will, anyway :)
greetings
Nico
2 thoughts. 1) I did try to make code that didn't work, but it DID work. 2) I also get…
<snip>
Again, we've been working on Linux for quite some time.
I'm sorry but don't these 2 statements contradict themselves. They weren't doing anything because they didn't have the resources but they have been doing things for quite some time. Huh??
An if they were already so fire sure about it and had been looking into it for some time how come they are only putting out a survey now to ascertain interest?
And if they had been doing research all along wouldn't they know that there was interest? Yet they put the survey on a machine that was already 70% busy and 2 hours after the announcement it bogged down to nothing.
I don't know I'm still not convinced that this is totally real yet.
Just so you know, it is extremely unlikely that we will have native support for Paradox tables on Linux.
Since a great deal of Delphi programming uses Paradox I guess we can assume that it will NOT be able to directly port Windows apps. Interesting.
2 thoughts. 1) I did try to make code that didn't work, but it DID work. 2) I also get…
Hello,
I believe it when I see or at least hear an announcement that they are doing it. Job postings do not a project make.
I copy some quotes from John Kaster at the end of this message. The last one isn't related to the topic, but it shows how Kaster's posting are considered "the voice of Borland" in their newsgroups.
yet but when it is done it will be done in such a way that any API can use it.
I finally succeeded with such API independance... just to realize that it would be messy for components developers :-(
Oh well, good luck with whatever you intend on doing. :-)
Thanks!! :-)
greetings
Nico
P.D. I will unsubscribe tonight, if you have any comments to me:
2 thoughts. 1) I did try to make code that didn't work, but it DID work. 2) I also get…
Nicolas Aragon wroteI've been thinking on what a Borland's Delphi for Linux will mean and I've changed my mind about the best way of creating a RAD for FPC on Linux, now that future Borland's product will be the one to beat.
I believe it when I see or at least hear an announcement that they are doing it. Job postings do not a project make.
Since I don't intend to use gtk anymore, I won't be able to help Lazarus in any way, so it makes no sense to stay messing around here.
Since the current layout we are doing for Lazarus is totally API independent it means that you don't have to use will be possible to use everything we are doing in Lazarus with Qt or something else right away. Of course we don't have the basic class libraries setup yet but when it is done it will be done in such a way that any API can use it.
Oh well, good luck with whatever you intend on doing. :-)