Sign in Join

Live chat for Free Pascal, Lazarus, Delphi and every Pascal programmer.

Mailing lists / lazarus / 2014-02

Lazarus compilation problems with 2.6.4 RC1

23 messages · 11 people · –

Reply by email Subscribe to this list Your subscription

Read-only archive. Use j and k to move between messages.

Michael Van Canneyt

Hi,

I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release candidate.

It failed on the sqldb package.
The sources mistakenly assume that the MySQL 5.6 connector will be included with 2.6.4. I fixed the define that controls this.

It then failed on the debugger package. I could not immediatly determine what the problem was, the unit with types was not found (when the sources were in fact there)

So, I first removed all .ppu and .o files in the lazarus dirs. It didn't help.

I then tried compiling on the command-line, using a make clean all.

This failed:

/bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled /bin/cp: cannot stat `Makefile.compiled': No such file or directory make[1]: *** [compiled] Error 1
make[1]: Leaving directory `/home/michael/projects/lazarus/components/debuggerintf' make: *** [debuggerintf] Error 2

I faked the file:

araminta: >cd /home/michael/projects/lazarus/components/debuggerintf araminta: >touch Makefile.compiled
araminta: >make PP=ppcx64-2.6.4
/bin/rm -f lib/x86_64-linux/debuggerintf.ppu
ppcx64-2.6.4 -MObjFPC -Scghi -O1 -g -gl -vewnhi -l -Fu../../packager/units/x86_64-linux -Fu../lazutils/lib/x86_64-linux -Fu. -Fu/usr/local/lib/fpc/2.6.4/units/x86_64-linux/rtl -FE. -FUlib/x86_64-linux -Cg -dx86_64 debuggerintf.pas /bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled

All went well, and I could build the IDE, and from then on I could build the IDE from inside the IDE.

Strange problems. So I decided to report here...

Michael.

Reply by email

patspiper
to Michael Van Canneyt
Michael Van Canneyt wrote 9 lines
Hi, I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release candidate. It failed on the sqldb package. The sources mistakenly assume that the MySQL 5.6 connector will be included with 2.6.4. I fixed the define that controls this.

Where exactly?

Show quoted text 35 lines
It then failed on the debugger package. I could not immediatly determine what the problem was, the unit with types was not found (when the sources were in fact there) So, I first removed all .ppu and .o files in the lazarus dirs. It didn't help. I then tried compiling on the command-line, using a make clean all. This failed: /bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled /bin/cp: cannot stat `Makefile.compiled': No such file or directory make[1]: *** [compiled] Error 1 make[1]: Leaving directory `/home/michael/projects/lazarus/components/debuggerintf' make: *** [debuggerintf] Error 2 I faked the file: araminta: >cd /home/michael/projects/lazarus/components/debuggerintf araminta: >touch Makefile.compiled araminta: >make PP=ppcx64-2.6.4 /bin/rm -f lib/x86_64-linux/debuggerintf.ppu ppcx64-2.6.4 -MObjFPC -Scghi -O1 -g -gl -vewnhi -l -Fu../../packager/units/x86_64-linux -Fu../lazutils/lib/x86_64-linux -Fu. -Fu/usr/local/lib/fpc/2.6.4/units/x86_64-linux/rtl -FE. -FUlib/x86_64-linux -Cg -dx86_64 debuggerintf.pas /bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled All went well, and I could build the IDE, and from then on I could build the IDE from inside the IDE. Strange problems. So I decided to report here...

Which Lazarus revision? I can compile 44074 without issues. The next few commits cause compilation to fail due to debugger related stuff, and not sqlsb stuff.

Is fpc 2.6.4 equivalent to rev. 26678?

Stephano

Reply by email

luiz americo pereira camara
to Michael Van Canneyt: I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release…

I'm having similar problem.

When trying to rebuild from the IDE, the compilation stops with the following message:
..\debugger\debugger.pp(41,3) Fatal: Can't find unit DbgIntfBaseTypes used by Debugger

I added debuggerintf dependency to Lazarus project and no luck Added debugger.pp to Lazarus project and also no luck

make clean all worked

Luiz

Luiz

2014-02-15 12:59 GMT-03:00 Michael Van Canneyt :

Show quoted text 53 lines
Hi, I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release candidate. It failed on the sqldb package. The sources mistakenly assume that the MySQL 5.6 connector will be included with 2.6.4. I fixed the define that controls this. It then failed on the debugger package. I could not immediatly determine what the problem was, the unit with types was not found (when the sources were in fact there) So, I first removed all .ppu and .o files in the lazarus dirs. It didn't help. I then tried compiling on the command-line, using a make clean all. This failed: /bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled /bin/cp: cannot stat `Makefile.compiled': No such file or directory make[1]: *** [compiled] Error 1 make[1]: Leaving directory `/home/michael/projects/lazarus/components/ debuggerintf' make: *** [debuggerintf] Error 2 I faked the file: araminta: >cd /home/michael/projects/lazarus/components/debuggerintf araminta: >touch Makefile.compiled araminta: >make PP=ppcx64-2.6.4 /bin/rm -f lib/x86_64-linux/debuggerintf.ppu ppcx64-2.6.4 -MObjFPC -Scghi -O1 -g -gl -vewnhi -l -Fu../../packager/units/x86_64-linux -Fu../lazutils/lib/x86_64-linux -Fu. -Fu/usr/local/lib/fpc/2.6.4/units/x86_64-linux/rtl -FE. -FUlib/x86_64-linux -Cg -dx86_64 debuggerintf.pas /bin/cp -f Makefile.compiled lib/x86_64-linux/DebuggerIntf.compiled All went well, and I could build the IDE, and from then on I could build the IDE from inside the IDE. Strange problems. So I decided to report here... Michael. -- _______________________________________________ Lazarus mailing list http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20140216/511d4a6b/attachment.html>

Reply by email

Maxim Ganetsky
to Michael Van Canneyt: I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release…

15.02.2014 19:59, Michael Van Canneyt ?????:

Show quoted text 8 lines
Hi, I tried to compile a new lazarus version from inside the IDE, using the 2.6.4 release candidate. It failed on the sqldb package. The sources mistakenly assume that the MySQL 5.6 connector will be included with 2.6.4.

It is included in current 2.6.3, but indeed missing in 2.6.4 RC1. Nobody tests compilation with 2.6.4 RC1, so I wouldn't recommend to use it.

Best regards, Maxim Ganetsky mailto:

Reply by email

Sven Barth
to Maxim Ganetsky
Maxim Ganetsky wrote 13 lines
15.02.2014 19:59, Michael Van Canneyt ?????: > > Hi, > > I tried to compile a new lazarus version from inside the IDE, using the > 2.6.4 release candidate. > > It failed on the sqldb package. The sources mistakenly assume that the > MySQL 5.6 connector will be included with 2.6.4. It is included in current 2.6.3, but indeed missing in 2.6.4 RC1. Nobody tests compilation with 2.6.4 RC1, so I wouldn't recommend to use it.

It's missing in 2.6.4 RC1, because MySQL 5.6 support was added after RC1 was tagged. 2.6.4 final will contain MySQL 5.6 support.

Regards,
Sven

Reply by email

Corpsman
to Sven Barth

Hi,

i have the same error using the FPC 2.6.2-5 compiler, is there a .deb file for the 2.6.3 compiler, i did not find any ?

Regards

Corpsman

Sven Barth wrote 26 lines
On 16.02.2014 14:12, Maxim Ganetsky wrote: > 15.02.2014 19:59, Michael Van Canneyt ?????: >> >> Hi, >> >> I tried to compile a new lazarus version from inside the IDE, using the >> 2.6.4 release candidate. >> >> It failed on the sqldb package. The sources mistakenly assume that the >> MySQL 5.6 connector will be included with 2.6.4. > > It is included in current 2.6.3, but indeed missing in 2.6.4 RC1. Nobody > tests compilation with 2.6.4 RC1, so I wouldn't recommend to use it. > It's missing in 2.6.4 RC1, because MySQL 5.6 support was added after RC1 was tagged. 2.6.4 final will contain MySQL 5.6 support. Regards, Sven -- _______________________________________________ Lazarus mailing list http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
Auf meiner Homepage www.Corpsman.de ist immer was los, st?ndig wird sie aktualisiert und erweitert. Da ist f?r jeden was dabei.

Reply by email

Mattias Gaertner
to luiz americo pereira camara: I'm having similar problem. When trying to rebuild from the IDE, the compilation stops…
Show quoted text 9 lines
I'm having similar problem. When trying to rebuild from the IDE, the compilation stops with the following message: ..\debugger\debugger.pp(41,3) Fatal: Can't find unit DbgIntfBaseTypes used by Debugger I added debuggerintf dependency to Lazarus project and no luck Added debugger.pp to Lazarus project and also no luck

The ide/lazarus.lpi is only for debugging the IDE.

make clean all worked

"make clean all" should create an IDE that can rebuild itself.

Mattias

Reply by email

Mattias Gaertner
to Sven Barth: It's missing in 2.6.4 RC1, because MySQL 5.6 support was added after RC1 was tagged.…
On 16.02.2014 14:12, Maxim Ganetsky wrote: [...] It's missing in 2.6.4 RC1, because MySQL 5.6 support was added after RC1 was tagged. 2.6.4 final will contain MySQL 5.6 support.

Both RC1 and final have 2.6.4 and can not be distinguished via IFDEFs. So I guess

{$IF FPC_FULLVERSION>= 20603}
{$DEFINE HASMYSQL56CONNECTION}
{$ENDIF}

is correct in future. Right?

Mattias

Reply by email

Sven Barth
to Mattias Gaertner
Mattias Gaertner wrote 16 lines
On Sun, 16 Feb 2014 14:19:22 +0100 Sven Barth wrote: > On 16.02.2014 14:12, Maxim Ganetsky wrote: > [...] > It's missing in 2.6.4 RC1, because MySQL 5.6 support was added after RC1 > was tagged. 2.6.4 final will contain MySQL 5.6 support. Both RC1 and final have 2.6.4 and can not be distinguished via IFDEFs. So I guess {$IF FPC_FULLVERSION>= 20603} {$DEFINE HASMYSQL56CONNECTION} {$ENDIF} is correct in future. Right?

Yes.

Regards,
Sven

Reply by email

Sven Barth
to Corpsman: i have the same error using the FPC 2.6.2-5 compiler, is there a .deb file for the 2.6.3…
Corpsman wrote
Hi, i have the same error using the FPC 2.6.2-5 compiler, is there a .deb file for the 2.6.3 compiler, i did not find any ?

FPC 2.6.2 does not come with MySQL 5.6 support so that's clear that there is an error. There are AFAIK no .deb-packages for snapshots. You'll need to wait until 2.6.4 is released which will be somewhen in the near future (though I don't have an exact date :) ).

Regards,
Sven

Reply by email

Bart
to Bart
Bart wrote
We were politely asked to test 2.6.4RC1, so we did. And we report these errors...

Just found out that I could not build any program (in Lazarus) with target win64.
"registerfcl.pas(43,3) Fatal: Can not find unit ServiceManager used by RegisterFCL. Check if package FCL is in the dependencies." (FCL is in dependencies)

Bart

Reply by email

Bart
to Bart
Bart wrote 5 lines
Just found out that I could not build any program (in Lazarus) with target win64. "registerfcl.pas(43,3) Fatal: Can not find unit ServiceManager used by RegisterFCL. Check if package FCL is in the dependencies." (FCL is in dependencies)

Does this look OK to you (compiler options)?

-MObjFPC -Scghi -Cirot -O1 -Twin64 -Px86_64 -g -gl -gh -vewnhiut -Filib\x86_64-win64 -FuMijnLib
-FuC:\devel\lazarus\components\printers\lib\x86_64-win64\win32 -FuC:\devel\lazarus\components\synedit\units\x86_64-win64\win32 -FuC:\devel\lazarus\components\lazcontrols\lib\x86_64-win64\win32 -FuC:\devel\lazarus\components\cairocanvas\lib\x86_64-win64\win32 -FuC:\devel\lazarus\lcl\units\x86_64-win64\win32
-FuC:\devel\lazarus\lcl\units\x86_64-win64
-FuC:\devel\lazarus\components\lazutils\lib\x86_64-win64 -FuC:\devel\lazarus\packager\units\x86_64-win64 -Fu. -FUlib\x86_64-win64 -l -dLCL -dLCLwin32

Somehow it appends win2 to several paths and it defines LCLWin32?

As a comparison the same project with target WinCE

-MObjFPC -Scghi -Cirot -O1 -Twince -Parm -g -gl -gh -vewnhi -Filib\arm-wince -FuMijnLib
-FuC:\devel\lazarus\components\printers\lib\arm-wince\wince -FuC:\devel\lazarus\components\synedit\units\arm-wince\wince -FuC:\devel\lazarus\components\lazcontrols\lib\arm-wince\wince -FuC:\devel\lazarus\components\cairocanvas\lib\arm-wince\wince -FuC:\devel\lazarus\lcl\units\arm-wince\wince
-FuC:\devel\lazarus\lcl\units\arm-wince
-FuC:\devel\lazarus\components\lazutils\lib\arm-wince -FuC:\devel\lazarus\packager\units\arm-wince -Fu. -FUlib\arm-wince -l -dLCL -dLCLwince

Bart

Reply by email

Mattias Gaertner
to Bart
Show quoted text 15 lines
On 2/16/14, Bart wrote: [...] -MObjFPC -Scghi -Cirot -O1 -Twin64 -Px86_64 -g -gl -gh -vewnhiut -Filib\x86_64-win64 -FuMijnLib -FuC:\devel\lazarus\components\printers\lib\x86_64-win64\win32 -FuC:\devel\lazarus\components\synedit\units\x86_64-win64\win32 -FuC:\devel\lazarus\components\lazcontrols\lib\x86_64-win64\win32 -FuC:\devel\lazarus\components\cairocanvas\lib\x86_64-win64\win32 -FuC:\devel\lazarus\lcl\units\x86_64-win64\win32 -FuC:\devel\lazarus\lcl\units\x86_64-win64 -FuC:\devel\lazarus\components\lazutils\lib\x86_64-win64 -FuC:\devel\lazarus\packager\units\x86_64-win64 -Fu. -FUlib\x86_64-win64 -l -dLCL -dLCLwin32 Somehow it appends win2 to several paths and it defines LCLWin32?

The "win32" is the widgetset. The win32 widgetset supports 64bit too.
It should be renamed. Maybe "winapi"?

Mattias

Reply by email

Mattias Gaertner
to Bart: Does this look OK to you (compiler options)? -MObjFPC -Scghi -Cirot -O1 -Twin64 -Px86_64…
Show quoted text 7 lines
On 2/16/14, Bart wrote: > Just found out that I could not build any program (in Lazarus) with > target win64. > "registerfcl.pas(43,3) Fatal: Can not find unit ServiceManager used by > RegisterFCL. Check if package FCL is in the dependencies." > (FCL is in dependencies)

Is the unit servicemanager only for win32, or is this a bug in the fpc make install?

Mattias

Reply by email

Michael Van Canneyt
to Mattias Gaertner
Mattias Gaertner wrote 13 lines
On Sun, 16 Feb 2014 22:37:49 +0100 Bart wrote: > On 2/16/14, Bart wrote: > >> Just found out that I could not build any program (in Lazarus) with >> target win64. >> "registerfcl.pas(43,3) Fatal: Can not find unit ServiceManager used by >> RegisterFCL. Check if package FCL is in the dependencies." >> (FCL is in dependencies) Is the unit servicemanager only for win32, or is this a bug in the fpc make install?

Barring bugs, it is also for win64.
So it is probably a bug in the fpc make install.

I will check this.

Michael.

Reply by email

Mark Morgan Lloyd
to Bart
Bart wrote 9 lines
On 2/17/14, Mattias Gaertner wrote: > The "win32" is the widgetset. The win32 widgetset supports > 64bit too. I understand, but the naming indeed is unfortunate > It should be renamed. Maybe "winapi"? win32-64?

By analogy with GTK, Qt etc.: if the API is substantially the same for the x86 and x86-64 targets, I think "polluting" the widget set with the CPU type would be undesirable.

I've got an SGI Itanium (IA-64) system here, it can generally run x86 binaries using some form of OS-level emulation, and /if/ I built a program for it using Lazarus I'd expect to be able to use the existing Windows widgetset rather than having to find an hypothetical one called something like win32-Itanic.

Mark Morgan Lloyd markMLl .AT. telemetry.co .DOT. uk [Opinions above are the author's, not those of his employers or colleagues]

Reply by email

Reinier Olislagers
to Mark Morgan Lloyd
Mark Morgan Lloyd wrote 21 lines
Bart wrote: > On 2/17/14, Mattias Gaertner wrote: > >> The "win32" is the widgetset. The win32 widgetset supports >> 64bit too. > I understand, but the naming indeed is unfortunate > >> It should be renamed. Maybe "winapi"? > > win32-64? By analogy with GTK, Qt etc.: if the API is substantially the same for the x86 and x86-64 targets, I think "polluting" the widget set with the CPU type would be undesirable. I've got an SGI Itanium (IA-64) system here, it can generally run x86 binaries using some form of OS-level emulation, and /if/ I built a program for it using Lazarus I'd expect to be able to use the existing Windows widgetset rather than having to find an hypothetical one called something like win32-Itanic.

Agreed. Winapi sounds good.

Reply by email

Jy V
to Reinier Olislagers

after more than 90 minutes of running FPCUP was cyclcing so I decided to launch Lazarus,
then the script FPCUP stopped and displayed:
TUniversalInstaller: building module anchodocking.... fpcup finished

Lazarus is by default setup with the anchordockingdsgn 0.5 -> this is perfect if it can stays this way to resemble Delphi as much as possible, the Lazarus IDE displays in the about box:
version 1.3
FPC version 2.7.1
SVN revision 44120
arm-linux-gtk2

Building a simple LCL project (I mean the default VCL project), displays warning messages:
crtbegin.o not found, this will probably cause a linking failure, crtend.o not found, this will probably cause a linking failure,

run OK,
Breakpoint, Debug OK,

I will try tomorrow to fossil xmlrad (meaning in your world Git/SVN pull XMLRAD project),
and I will keep reporting as I progress,
but at this stage it looks that you have provided 1 way to install Lazarus for ARM,

however, the procedure requires multiple steps,
so I would recommend that you provide a single shell script to download, and this shell script would perform all the intermediate steps that you have written in the wiki page,

Regards, Jerome.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20140217/510054be/attachment-0001.html>

Reply by email

Bart
to Jy V
Jy V wrote
after more than 90 minutes of running FPCUP was cyclcing so I decided to launch Lazarus,

This thread is about fpc 2.6.4RC1 i.r.t. Lazarus.
Please do not hijack this thread to discuss fpcup or other stuff.

Bart

Reply by email

Reinier Olislagers
to Bart: We were politely asked to test 2.6.4RC1, so we did. And we report these errors... If we…
Bart wrote 8 lines
On 2/16/14, Maxim Ganetsky wrote: > ... Nobody > tests compilation with 2.6.4 RC1, so I wouldn't recommend to use it. We were politely asked to test 2.6.4RC1, so we did. And we report these errors... If we wouldn't have, the what?

AFAIU, Marco commits fixes to 2.6.4RC1 to the fixes 2.6 branch, and there have been some fixes - so testing with fixes 2.6 would seem to make more sense at this time.

Marco, is this accurate?

Reply by email