gmane.comp.compilers.free-pascal.general

Re: TList Notify count changed

Zaher Dirkey — 2012-05-23T13:02:59



Might i ask to make TList.SetCount a virtual method?

Best Regards
Zaher Dirkey
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Re: 2.6.0 for Solaris? And other questions[oneresolved]

2012-05-23T09:28:31

Thanks Mark. I haven't gotten to my SPARC boxes yet. Since I already had a copy of 2.6.0 built on my Linux box I was playing around with building there first. 

-----Original Message-----
From: Mark Morgan Lloyd <markMLl.fpc-pascal< at >telemetry.co.uk>
Sender: fpc-pascal-bounces< at >lists.freepascal.org
Date: Wed, 23 May 2012 09:05:55 
To: <fpc-pascal< at >lists.freepascal.org>
Reply-To: FPC-Pascal users discussions <fpc-pascal< at >lists.freepascal.org>
Subject: Re: [fpc-pascal] Re: 2.6.0 for Solaris? And other questions [one
resolved]

microcode< at >zoho.com wrote:

That will require that you can find a precompiled libgdb- preferably a 
version that's already supported by the options in gdbint.pp. Or build 
one from source (which I've done on several systems).

I've just checked my SPARC Solaris 10 and I've not gone to the trouble: 
fp was built without gdblib.

You normally have to set up the ppc symlink manually, irrespective or 
platform. I usually use two stages, e.g. ppcsparc -> ppcsparc-2.4.4 and 
ppcsparc-2.4.4 -> /usr/local/lib/fpc/2.4.4/ppcsparc, and in cases where 
I'm e.g. running Lazarus I tell it to use the "one in the middle" i.e. 
ppcsparc-2.4.4.

Re: Re: 2.6.0 for Solaris? And other questions [oneresolved]

Mark Morgan Lloyd — 2012-05-23T09:05:55

That will require that you can find a precompiled libgdb- preferably a 
version that's already supported by the options in gdbint.pp. Or build 
one from source (which I've done on several systems).

I've just checked my SPARC Solaris 10 and I've not gone to the trouble: 
fp was built without gdblib.


You normally have to set up the ppc symlink manually, irrespective or 
platform. I usually use two stages, e.g. ppcsparc -> ppcsparc-2.4.4 and 
ppcsparc-2.4.4 -> /usr/local/lib/fpc/2.4.4/ppcsparc, and in cases where 
I'm e.g. running Lazarus I tell it to use the "one in the middle" i.e. 
ppcsparc-2.4.4.

TList Notify count changed

Zaher Dirkey — 2012-05-22T18:08:16

In TList how can i notify if Count is changed with New and Old values as
like in SetCount, SetCount procedure not virtual?
procedure SetCount(NewCount: Integer);

I am using last revision of Lazarus, FPC 2.6

Best Regards

Zaher Dirkey
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Jürgen Hestermann — 2012-05-22T16:29:22

Jonas Maebe schrieb:
 > That's a non sequitur:
 > a) even though the original code was written a long time ago, it has 
been changed many times since then
 > b) if it isn't broken, don't fix it. Example: changing all 
occurrences of "string" into "shortstring" will break the system unit.
 > The reason is that a "var a: string" is an "openstring" parameter 
(with a hidden "high" parameter that tells you the size of
 > the variable that was passed in), while "var a: shortstring" is not. 
So at the very least, you'd have to change var/out
 > parameters into openstring rather than into shortstring.

As far as I know you can pass ShortString of any size to a ShortString 
of maximum length (255). So I don't see how "var a : ShortString" would 
lead to errors. But of course, you need to be careful.

Nevertheless, I would rewrite such code. Even at the danger of breaking 
something else (which would hopefully appear directly) I would consider 
clear code so important that I would not rely on some compiler switch 
(that may change some day) to make it run. Especially because *if* the 
compiler switch is changed it would not strike you immediately. It would 
run a long time until someone (by accident) has a path longer than 255 
characters. And then it would not raise an exception (which would point 
you to the cause) but it would overwrite other data and then crash at a 
completely unrelated instruction (or just lead to wrong results).

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Re: 2.6.0 for Solaris? And other questions [oneresolved]

2012-05-22T10:22:08

On Tue, 22 May 2012 10:29:28 +0200 Sven Barth <pascaldragon-gM/Ye1E23mwN+BqQ9rBEUg< at >public.gmane.org>
wrote:


That's a good explanation, thank you. I would like to have the debugger
support. 

I'm planning to go through the build again a few times for various reasons
including I want to get a current copy on Solaris SPARC. I didn't get (all)
the units built and there are some other minor annoyances in the way it
worked out (no symlink from fpc to the ppcx module) and if it doesn't go
like you said I'll post back to the list.

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Re: 2.6.0 for Solaris? And other questions [oneresolved]

Sven Barth — 2012-05-22T08:29:28

Am 21.05.2012 14:59, schrieb microcode-ytc+IHgoah0< at >public.gmane.org:

This might be, because you compiled it without a libgdb which in turn is 
the one who links dynamically to libc. Most FPC applications (if they 
don't use threads or a WideString manager) don't require libc and thus 
don't need to link to it (which AFAIK is also the case for the IDE if it 
doesn't link to libgdb).

Regards,
Sven

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Jonas Maebe — 2012-05-22T08:22:50

Jürgen Hestermann wrote on Tue, 22 May 2012:


That's indeed probably true.


That's a non sequitur:
a) even though the original code was written a long time ago, it has  
been changed many times since then
b) if it isn't broken, don't fix it. Example: changing all occurrences  
of "string" into "shortstring" will break the system unit. The reason  
is that a "var a: string" is an "openstring" parameter (with a hidden  
"high" parameter that tells you the size of the variable that was  
passed in), while "var a: shortstring" is not. So at the very least,  
you'd have to change var/out parameters into openstring rather than  
into shortstring.

Adding support for new functionality (such as unicode file names)  
would obviously require revising all of that code, but that's because  
the context changes rather than because it's old.


Jonas
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: 2.6.0 for Solaris? And other questions [oneresolved]

2012-05-21T12:59:11


Hooray! See attached. Thanks Tomas!

Interestingly enough, the build process I stumbled through created
statically linked executables. In 2.6.0 fp was dynamically linked and
pointed to a copy of glibc I didn't have. In the 2.7.1 I created today, fp
is static and has no dependency on a particular version of anything. I
wonder if it could be created that way generically in the regular build
process and then everybody on Linux could run it.


Oops, I knew I forgot something! :-( Got the messages and didn't take the
time to figure out where to get the library and forget about your email
until I read it again now.


Paying a little more attention now after the build, the problem I am having
is not finding many of the units required to build the examples. I see I do
have quite a few libraries in the units directory but obviously I am
missing others. I'll look at this more later on.

Thanks to everyone who's helped, I seem to be getting a lot further and fp
is now working (without the debugger anyway, my fault) and looks great!
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Jürgen Hestermann — 2012-05-22T05:15:51


Sven Barth schrieb:
That would be explanation.
But then it must be *very* old code which shows that a revise is overdue.

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

You can probably ignore this

2012-05-22T05:05:41

It seems I have to post twice or nothing shows up to the list!

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Can't install FPC on mac bookair 10.7.4

Jonas Maebe — 2012-05-21T08:51:18

Chovav Sutachy wrote on Mon, 21 May 2012:


First, install Xcode if you do not have it yet. If you installed Xcode  
4.3 or later, next open Xcode, go to Preferences -> Downloads and  
install the "Command line tools" package (the installer doesn't know  
about this limitation because Xcode 4.3 was released after FPC 2.6.0).  
Then you will be able to install FPC.


Jonas
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Can't install FPC on mac bookair 10.7.4

Chovav Sutachy — 2012-05-21T05:54:45

Dear

I can't install FPC compiler on mac 10.7.3

Please see attachment files

Thank you

Chovav

P.S. -http://www.lazarus.freepascal.org/index.php/topic,16125.0.html




On Sun, May 20, 2012 at 5:00 PM, <fpc-pascal-request-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org>wrote:

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Sven Barth — 2012-05-21T06:57:34

Am 21.05.2012 07:17 schrieb "Jürgen Hestermann" <juergen.hestermann< at >gmx.de>:
$H+.
ShortString.
never change
and also

Estimated guess: the code is from a time where "String = AnsiString" wasn't
possible.

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Jürgen Hestermann — 2012-05-21T05:17:29


Tomas Hajny schrieb:
Ok. If that can be guaranteed.
But I am always wondering why in such situations the type is not set to 
ShortString.
Why use generic String type if all code relies on the fact that it may 
never change
from ShortString to some other string type? It's just asking for trouble 
and also
makes the code less clear.

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Darwin i386

Andrew Brunner — 2012-05-20T15:00:07

On Sun, May 20, 2012 at 8:52 AM, Jonas Maebe <jonas.maebe-3RqwKoEL1alVsUKGZV2E3Q< at >public.gmane.org> The
above was a question relating to "... with the exception of threads calling
methods via synchronize (so I can debug)." I.e., what are the exact things
you are doing, what is happening, and what are the expected results
regarding trying to debug threads calling methods via synchronize (which, I
assume, works on other platforms but not on Mac OS X).

Sorry. Kinda missed that. Debugging heavily threaded code under linux has
proved to be difficult. So I have developed a directive that triggers
threaded code to execute via synchronize.

I recently cleared all linux issues with a core networking engine (both
client and server) and moved to Apple to test. On Apple I needed to step
into code that normally runs reentrant from the main thread, from other
threads, and from other threads via synchronize.

On Linux all was fine. On Apple - I was seeing thread locking to my
network engine in client mode. Same code, different operation system. I
traced all the way to the signaling mechanisms in the classes unit. The
thread went into wait infinite and never executed the method.

So instead of complaining about it here - I decided to method scheduling
system for the engine. This way when the engine is idle it can execute
methods in its own thread space. During debug mode I run engine
callbacks via synchronize. On linux I get no problems. On Apple I had
gotten tons of timeout errors and weird behavior that takes 5-10 seconds to
step over or into each line!

So I have gdb 7.4.1 from source installed and I can't test. I may have to
revert back to gdb stock just to get back to work :)
_______________________________________________
fpc-pascal maillist - fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Darwin i386

Jonas Maebe — 2012-05-20T13:52:23

On 20 May 2012, at 15:25, Andrew Brunner wrote:

The above was a question relating to "... with the exception of threads calling methods via synchronize (so I can debug)." I.e., what are the exact things you are doing, what is happening, and what are the expected results regarding trying to debug threads calling methods via synchronize (which, I assume, works on other platforms but not on Mac OS X).

Mainline GDB 7.1 is not supported on Mac OS X 10.7. The first GDB version with 10.7 support is GDB 7.4 (see the release notes at http://sourceware.org/gdb/wiki/GDB_7.4_Release, search for "lion")

The above message means that you have not regenerated the .dSYM bundle with debug information for your application via passing -Xg to the compiler command line options. Either delete the .dSYM bundle, or recompile with -Xg or run dsymutil on the generated binary (make sure to select DWARF debug information when using -Xg/dsymutil on Mac OS X, with Stabs it will probably disable debugging instead).

That is unrelated to FPC or even Lazarus. Maybe mainline GDB does not take into account the property list of the app bundle to launch the application as a foreground application. You can try to work around that by using TransformProcessType with the kProcessTransformToForegroundApplication parameter in your program's initialization code (see e.g. http://cocoadev.com/wiki/TransformProcessType )

You don't have to recompile FPC, the default FPC installation already includes a 64 bit cross-compiler ("ppcx64", or "fpc -P x86_64"). To compile a 64 bit Lazarus on Mac OS X, you will have to select the Qt widgetset since Carbon is not supported in 64 bit mode (and the Cocoa widegetset is not yet sufficiently far along). Regarding how to compile Lazarus so it uses the Qt widget set: I have no idea how to do that, please ask on the Lazarus mailing list (or search the wiki with Google, it's probably documented there somewhere as well).

Jonas_______________________________________________
fpc-pascal maillist - fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Darwin i386

Andrew Brunner — 2012-05-20T13:25:20

I'm upgrading my gdb from shipping version with darwin to gdb version from
download.  I downloaded and ran /configure make make install 7.1.  It
doesn't work with fpc.  I'm going to have to switch back.  The newer
versions require code signing...

warning: dsym file UUID doesn't match the one in
/Developer/Source/Builds/Aurawin/AuCloud/Apple/32/AuCloud
(no debugging symbols found)...done.
(gdb) run
Starting program: /Developer/Source/Builds/Aurawin/AuCloud/Apple/32/AuCloud
Unable to find Mach task port for process-id 62916: (os/kern) failure (0x5).
 (please check gdb is codesigned - see taskgated(8))
(gdb)

(gdb) run
Starting program: /Developer/Source/Builds/Aurawin/AuCloud/Apple/32/AuCloud
Unable to find Mach task port for process-id 62916: (os/kern) failure (0x5).
 (please check gdb is codesigned - see taskgated(8))
(gdb)

BUT: sudo gdb bypasses this problem, and the program can get past this
error but when the program launches, gdb gives

(gdb) run
Starting program: /Developer/Source/Builds/Aurawin/AuCloud/Apple/32/AuCloud
[New Thread 0x2003 of process 62922]
[New Thread 0x2103 of process 62922]
[New Thread 0x2203 of process 62922]
[New Thread 0x2303 of process 62922]
[New Thread 0x2403 of process 62922]
[New Thread 0x2503 of process 62922]
[New Thread 0x2603 of process 62922]
[New Thread 0x2703 of process 62922]
TCarbonTabsControl.SetFocus Error: SetKeyboardFocus  failed with result
-30585
warning: dsym file UUID doesn't match the one in
/Developer/Source/Builds/Aurawin/AuCloud/Apple/32/AuCloud
(no debugging symbols found)...done.

BUT notice TCarbonTabsControl.SetFocus Error: SetKeyboardFocus  failed with
result -30585.  Now the main screen has no focus.  I can only close the
window :-)

Progress but I need to get gdb 7.1 to run properly on darwin.  Either that
or I gotta restore the old gdb :-(

Google "gdb" together the exact error message you are getting, and you will
I've been.  I didn't have a problem with GCC or GDB.  I've got the latest
and greatest of both.  This is a darwin thing.  gdb can't operate properly
without root access.

I've built a system cert.  I've used codesign -s  gdb-cert ./gdb and darwin
asked for my admin creds.  So I know gdb has been signed.
I used which gdb to make sure I was in the right place to codesign.

Can you tell me how to compile fpc to make 64bit lazarus?
_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Tomas Hajny — 2012-05-20T13:06:56

No, unit System (which uses these include files) is never compiled with $H+.

Tomas


_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

Re: Unicodestrings and Assign(File)

Jürgen Hestermann — 2012-05-20T12:40:53

In "fpc\rtl\inc\file.inc" (and fpc\fpc\rtl\inc\text.inc) the declaration 
of "Assign"
uses the generic "String" type for the file name. I think this can be
AnsiString too, so it may be longer than 255 characters.
But the assignment to the internal *rec element "name" (which
is "array[0..255] of char") is done with

Move(s[1],TextRec(t).Name,Length(s));

That would be a bug IMO when S is longer than 255 characters.
"Move" does no range check.

_______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

FPMake configure script

2012-05-20T11:22:11

  

Hi, 

I've been playing with a configuration script for my project
some time now. What started as a simple compiled application, it has
changed into a instantFPC script. Today I finally refactored the
majority of the code into a set of classes very similar to how fpmake
works. 

I've attached an example config script together with the utils
unit. My plan is to add some tool classes in the future that check for
installed libraries etc. Please note that this is a very basic
implementation and I'm open for suggestions and tips. Ultimately I would
like to see this to be included in FPC trunk if everyone agrees.


Regards, Darius 
  _______________________________________________
fpc-pascal maillist  -  fpc-pascal-PD4FTy7X32k2wBtHl531yWD2FQJk+8+b< at >public.gmane.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal