[fpc-pascal] Win64 release of FPC 2.6.2 missing

Jonas Maebe jonas.maebe at elis.ugent.be
Tue May 7 08:33:42 CEST 2013

On 06 May 2013, at 18:40, Graeme Geldenhuys wrote:

> On 2013-05-06 08:32, Jonas Maebe wrote:
>> The whole point is that the "typical end user" doesn't know nor should  
>> have to care about whether the compiler is a 32 or 64 bit binary. It  
>> should not make any difference whatsoever.
> How does the saying go: "I want to live in Theory, because everything
> works in theory." ;-)

No, it goes: "nobody ever submitted a bug report about this nor posted anything on the FPC mailing list and I don't use Windows myself, so this is the first I hear about it". Problems need to be reported and fixed.

> And since when is a developer a "typical end user"?

That's a non sequitur. As mentioned, it has to work out of the box.

> If I run 64-bit Mac or 64-bit Windows or 64-bit Linux, I want 64-bit
> applications. Some OS's allow for easier execution of 32-bit apps, but
> definitely not all.

That's why I explicitly mentioned in my previous mail that the thread was about Windows. We do offer 64 bit compiler binaries for Linux.

> All my systems run 64-bit OS's, and all my installed
> libraries etc are 64-bit too (think Firebird, etc). 32-bit executables
> cannot load 64-bit libraries, so I need 64-bit executables.

That is about the produced programs, not about the compiler binary itself.

> To put it simply, that is the reason I (and many others) need a 64-bit
> FPC. I've had lots of bad experiences with FPC cross-compilers - from
> installing headaches, to broken executables (don't ask me why, I don't
> know why). So I simply don't trust FPC cross-compliers. This is why I
> have various build systems in VM's for all our supported platforms.

Every single person using the Mac OS X version to build 64 bit applications uses a cross-compiler and I've never seen any bug reports or complaints about it. As mentioned before, I don't remember seeing any bug reports either related to the Win64 compiler being a cross-compiler (other than people asking to get a 64 bit binary with as only reason that it would then be a 64 bit binary).

I'm not saying that you are wrong when you say that you have had problems with (self-built?) cross-compilers (where are the bug reports?), just that this is all unrelated to an official installer provided by us that is fully tested and supported (other than the Lazarus integration problem that apparently exists), and that there are plenty of data points to indicate that cross-compiling with FPC in general is not intrinsically broken nor hard (cross-building FPC yourself can be harder, but as you mentioned we are talking about typical end-users and not about developers).

> And just to make it clear... it was NOT my idea to release my build of
> FPC 2.6.4 for 64-bit Windows. I was asked in private to archive what I
> have and make it available for download (on my own server). That is what
> I did.

I don't mind that you did, only the fact that a personal installation was placed in the distribution location on the official FPC ftp site (and I know you are not responsible for that in any way).


More information about the fpc-pascal mailing list