Page layout differs between pages in multipage report

[balazsszekely]

@Zoran
But for quite some time now Fortes cannot compile with Lazarus trunk (1.7) in Windows. See: https://sourceforge.net/p/fortes4lazarus/bugs/8/.

Somehow I missed your post. Support for win98 was droped with FPC 3.0.0 that is why you cannot compile Fortesreport. If you replace OpenPrinter with OpenPrinterW everything works fine, you can install fortres4lazarus.

[Zoran]

@Zoran
But for quite some time now Fortes cannot compile with Lazarus trunk (1.7) in Windows. See: https://sourceforge.net/p/fortes4lazarus/bugs/8/.

Somehow I missed your post. Support for win98 was droped with FPC 3.0.0 that is why you cannot compile Fortesreport. If you replace OpenPrinter with OpenPrinterW everything works fine, you can install fortres4lazarus.

Thank you, GetMem. I use Windows 7 and Windows 8, I do not need Win98 support. I will change, but still Fortes4Lazarus source should be updated to compile.

The FortesReport CE 4.0 version does compile with trunk.
I have Lazarus 1.7 (about a week old) running and FortesReport works well with it.
Have you tried the new Fortesreport CE 4.0 version? You can find it at: https://github.com/fortesinformatica/fortesreport-ce

Uh... 
What is this? Another version of Fortes??? Why another version now? Is it a new port, independent from LuizAmerico's?
What does "ce" stand for?

[Zoran]

The FortesReport CE 4.0 version does compile with trunk.
I have Lazarus 1.7 (about a week old) running and FortesReport works well with it.
Have you tried the new Fortesreport CE 4.0 version? You can find it at: https://github.com/fortesinformatica/fortesreport-ce

Uh... 
What is this? Another version of Fortes??? Why another version now? Is it a new port, independent from LuizAmerico's?
What does "ce" stand for?

And this "ce" version doesn't compile on my Linux Mint 64-bit:

RLXLSXFileFormat.pas(497,3) Error: Argument cannot be assigned to

It seems it's not OS problem, but a 64-bit issue - it tries to cast Pointer as Integer and then assign to it. So, for me, switching to this version seems even more dangerous - it might have serious compatibility issues which are hidden on 32-bit system.

[JanRoza]

CE stands for Community Edition and it does the job perfectly for me while running Windows 10 64-bit.
But I use the 32-bit version of Lazarus 1.7 so I can't tell you how it compiles using Lazarus 64-bit.

[Zoran]

CE stands for Community Edition and it does the job perfectly for me while running Windows 10 64-bit.
But I use the 32-bit version of Lazarus 1.7 so I can't tell you how it compiles using Lazarus 64-bit.

Thanks, "ce" confused me, I thought it might be for Windows ce.
Still, what is it? Why another port? Is Luiz Americo's port abandoned now?

Try to compile a project with Fortes dependency to 64 bit. It won't compile.
Although this line (unit RLXLSXFileFormat.pas, line 497) is easily fixable (should just change all "Integer" in this line to "PtrInt"), I wander how many such mistaken 32-bit Integers can be there, which are not in left-side of assignment, so will pass compile, but will fail once in runtime, and that is not easy to track (it might just work many times and once give who knows what kind of wrong behaviour).

[JanRoza]

Maybe Luiz can shed some light on this community edition.
As it has a different website it looks like an independent undertaking.
But since it works great I'm happy with it 

[Zoran]

Maybe Luiz can shed some light on this community edition.
As it has a different website it looks like an independent undertaking.
But since it works great I'm happy with it 

But once you need to compile for 64-bit, you will encounter serious issues. For now, 64-bit Windows can run 32-bit applications well, but in future you might need to compile 64-bit even for Windows. For 64-bit linuxes you can't run 32-bit, so this does not work. For me, for this reason, using this package is out of question.

I found this bug was reported already, so I made comment there: https://github.com/fortesinformatica/fortesreport-ce/issues/93#issuecomment-199384799

[JD]

@Zoran
The FortesReport CE version is listed as 4.0 so it appears to be the most recent
http://forum.lazarus.freepascal.org/index.php/topic,31413.msg201261.html#msg201261

Concerning the problem encountered using 64 bit Lazarus, you might need to contact elidorio directly.

JD

[Zoran]

@Zoran
The FortesReport CE version is listed as 4.0 so it appears to be the most recent
http://forum.lazarus.freepascal.org/index.php/topic,31413.msg201261.html#msg201261

Concerning the problem encountered using 64 bit Lazarus, you might need to contact elidorio directly.

JD

Elidorio filed the bug report to which I gave link in my previous post, but he does not seem to be a contributor of the package, see: https://github.com/fortesinformatica/fortesreport-ce/graphs/contributors

[jesusr]

I've defined a report with lazreport and this shows/pronts well if the report is only one page long.
As soon as there is more data than fits on one page the second page is missing some of the layout items from the first page (see attached screen print) Page 2 misses a number of lines (top and side lines).
Is this a bug or am I using lazreport incorrectly?

Yep, it's been a while since you report but anyway, if you want please submit a bug report and attach a sample report so this could be sorted out in future versions.

[jesusr]

I agree! Not just this line draw thing, generally speaking is more stable, the designer is more functional, etc. Try to run Lazreport on Mac, to be more specific export something to pdf, it's not gonna work.

Interesting, exporting to pdf has some limitations yes, but that is not limited to Mac. You say you don't use LazReport anymore but anyway do you have any sample that doesn't work? I would like to take a look.

[LuizAmérico]

The community edition is the original delphi repository that also compiles in Lazarus.

I'll take a look so all functionality in my port is present so the development can done only in that repository

Luiz

[Zoran]

Luiz fixed Fortes4Lazarus. Thank you, Luiz.

I'll take a look so all functionality in my port is present so the development can done only in that repository

Please don't abandon Fortes4Lazarus soon. At least not before Fortes-CE is well tested with 64-bit.
They still haven't fixed the obvious 64-bit issue. It was reported twice:
https://github.com/fortesinformatica/fortesreport-ce/issues/82 (opened 6. january)
https://github.com/fortesinformatica/fortesreport-ce/issues/93 (opened 19. february)

So, they don't seem to respond very promptly on reported bugs.
Okay, they will probably fix this soon, but then with Fortes-CE we will still be far from the stability we have with Fortes4Lazarus.
I'm afraid that this 64-bit issue can be much more dangerous (for the reasons I explained in my comment from 21. march):

You cannot compile Fortes on 64-bit target because you cannot cast pointer type (which is 64 bit) to Integer (which is 32-bit) without loss. And when it is on left side of the assignment this is compile-time error.
This is easily fixable by replacing Integer with PtrInt in both places in this line, that is:
PtrInt(ItemPtr) := PtrInt(FHeapPtr) + FHeapLength;
However, the biggest problem with this is that we can't be sure how many mixing Pointers and Integers are there, when they are not on left-side of the assignment - all these places will be much more difficult to find, because then the code will pass in compile time, but it will make serious issues once in runtime. The program may even work well for many times and once you get errors which are very hard to track.

And if they do not respond for so obvious compilation bug - how soon we can expect them to fix the who knows how many hard-to-spot 64-bit bugs?

On the other hand, your old port is proven to be stable - it works well for years without any intervention - until this recent change in LCL sources, you didn't need to fix anything since the small patch I gave you three years ago - see the history of changes - except some russian translation, no changes have been made since almost three years ago! And all have worked well until this change! I haven't encounter any problems on Linux and Windows, on 32-bit and 64-bit!

Therefore the maintenance of Fortes4Lazarus doesn't require too much work from you, so please keep maintaining it until the Fortes-CE team not only fix their code to compile on 64-bit, but until it passes the tests of time with Lazarus users (as your Fortes4Lazarus surely did).

[JanRoza]

@Luiz: I'll try to get a sample lazReport that shows the problem this weekend. Thanks for looking into this.

[LuizAmérico]

I'll keep the support for the old port. At the same time i will help the new port with the current and future issues.