FPC 3.0.0 "Pestering Peacock" lazarus / 60-Free Pascal ---------------------------------------------------------------------- marcov · 2015-11-26 19:19 Hello, We are happy to announce the release of the Free Pascal Compiler version 3.0.0 "Pestering Peacock". This version includes support for, a.o., several new platforms, codepage-aware strings, and an integrated Pascal source-repository (fppkg). New features are listed at http://wiki.freepascal.org/FPC_New_Features_3.0 Changes that may break backwards compatibility are documented at http://wiki.freepascal.org/User_Changes_3.0 The 3.0.0 docs is available at http://www.freepascal.org/docs-html/3.0.0/fpctoc.htm Downloads are available at our website: http://www.freepascal.org/download.var Enjoy! The Free Pascal Compiler Team                              Free Pascal Compiler                                 Version 3.0.0 ********************************************************************** ****** * Introduction ********************************************************************** ****** Please also read the platform specific README file, if it exists. This package contains an advanced freeware 32-bit/64-bit/16-bit pascal compiler for several platforms. The language and the runtime library are almost compatible with Turbo Pascal 7.0 and recent Delphi releases. There is also partial support for the Macintosh pascal dialect. News and changes related to this release are described in whatsnew.txt. Free Pascal 3.0.0 is currently available for the following platforms: - Linux-i386 - Linux-x86_64 (amd64) - Linux-powerpc - Linux-sparc - Linux-ARM - Win32-i386 (2000/XP, WinNT or later) - Win64-x86_64 (XP or later) - Wince-ARM (cross compiled from win32-i386) - FreeBSD-i386 - FreeBSD-x86_64 - Mac OS X/Darwin for PowerPC (32 and 64 bit) - Mac OS X/Darwin for Intel (32 and 64 bit) - iOS (ARM and AArch64/ARM64) and iPhoneSimulator (32 and 64 bit) - OS/2-i386 (OS/2 Warp v3.0, 4.0, WarpServer for e-Business and eComStation) - Haiku-i386 - GO32v2-i386 - Nintendo Gameboy Advance-ARM (cross compile from win32-i386) - Nintendo DS-ARM (cross compile from win32-i386) - Nintendo Wii-powerpc (cross compile from win32-i386) - AIX 5.3 and later for PowerPC (32 and 64 bit) - Java JVM (1.5 and later) and Android Dalvik (Android 4.0 and later) - Android (ARM, i386, MIPS) via cross-compiling. - MSDos-i8086 (cross compiled from win32-i386 or Linux) - Amiga, MorphOS and AROS There are other platforms (like other BSD variants) which are more or less working, but there is currently no maintainer for them and thus we cannot provide adequate support. If you want to change this and create and maintain versions for other platforms and targets, feel free to contact us, e-mail addresses are listed below. ********************************************************************** ****** * Features ********************************************************************** ****** - high speed compiler - fully 16, 32 or 64-bit code    - 32 bit compiler can be used on x86_64 Linux distributions to compile      32 bit applications - language features:    - almost fully compatible with Borland Pascal and Borland Delphi    - ansi strings    - wide strings    - exception support    - RTTI support    - procedure overloading    - generics (experimental)    - operator overloading    - COM, CORBA and raw interfaces support    - dynamic array support    - variant support    - inlining - code optimizer:    - peephole optimizer (80x86 only)    - jump optimizer    - loading of variables into registers    - assembler level dataflow analyzer (80x86 only)    - stack frame eliminations    - sophisticated register allocator - integrated BASM (built-in assembler) parser    - supports ATT syntax used by GNU C    - supports Intel syntax used by Turbo Pascal (x86-only) - can compile code into assembler source code for these assemblers:    - GNU Assembler (GAS)    - Netwide assembler (Nasm)    - Microsoft Assembler/Turbo Assembler (Masm/Tasm)    - Watcom assembler (wasm) - internal assembler for ultra fast object file generation - can call external C code    - h2pas utility to convert .h files to Pascal units - smart linking (also known as dead code stripping) - support for the GNU debugger - integrated development environment (disabled by default on Mac OS X)    - powerful user friendly Wordstar compatible multi file editor    - context sensitive help supports help files in HTML, Windows HLP and      Borland TPH format.    - debugger on most platforms - can create binaries running natively under both DOS and OS/2 (EMX version) - no need for Linux distribution specific binaries, programs you write run    on all distributions - high quality documentation ********************************************************************** ****** * Minimum requirements ********************************************************************** ****** Processor:    i386, x86_64, PowerPC, ARM, MIPS, i8086, Motorola 68k, Sparc or JVM    runtime Win32:   - MS Windows NT/2000/XP or later (Win95, 98 and ME are not supported     in FPC officially any longer).   - 16 MB RAM OS/2:   - OS/2 Warp v3.0 with one of late fixpaks - FP 35 should be fine,     OS/2 Warp v4.0 with FP 5 and above, WSeB, MCP or any eComStation version     (OS/2 2.0/2.1 currently not supported, not even for compiled programs).   - Certain features (e.g. Unicode support in the RTL) require proper support     only available in later OS/2 versions (at least OS/2 Warp 4.0 FP14),     otherwise the functionality may be limited (but the binaries still work     even on the older versions, albeit with some limitations). Linux:   - system running a 2.4.x kernel FreeBSD: - There are separate releases for FreeBSD 7.x system or 8.x system. The code      works on 6.x too. Older versions might need specific adaptations. Mac OS X:   - Mac OS X 10.4 and higher ---------------------------------------------------------------------- timppl · 2015-11-27 07:34 Thanks to all the developers. I now have it installed and one thing I have noticed is that when compiling from a terminal the output seems to scroll faster so I guess V3 is faster than 2.6.4  Regards Tim ---------------------------------------------------------------------- taazz · 2015-11-27 08:14 Congrats guys. An other excellent release. Thank you. ---------------------------------------------------------------------- hnb · 2015-11-27 08:59 Significant step forward . Now we need more detailed optional RTTI, anonymous methods and attributes ^^. Bye bye Delphi. ---------------------------------------------------------------------- Deepaak · 2015-11-27 16:24 Thank you everybody, who have given their valuable time and knowledge for this excellent release. ---------------------------------------------------------------------- Graeme · 2015-11-27 17:59 Very impressive list of features. Well done to everybody that was involved. ---------------------------------------------------------------------- Michl · 2015-11-27 18:28 Thank you so much for all you work!!! ---------------------------------------------------------------------- Blaazen · 2015-11-27 18:59 Thanks. BTW: "Pestering Peacock" is some Ubuntu related joke? ---------------------------------------------------------------------- marcov · 2015-11-27 19:10 Quote from: Blaazen on November 27, 2015, 07:59:18 pmThanks. BTW: "Pestering Peacock" is some Ubuntu related joke?  Not really. It was about two years ago during a FPC hackathon when it turned out the hotel where we staying had a bunch of peacocks. They were very noisy and annoying. During that meeting, the plans for 3.0 (and e.g. the unicode implementation of the RTL) were finalized, and sb proposed to call it "Pestering Peacock", and that stuck ---------------------------------------------------------------------- SnoopyDog · 2015-11-27 19:13 Thanks a lot for the new version I am using Lazarus 1.4.4 with FPC 2.6.4 currently. I tried to compile one of my projects where i use overloaded properties with FPC 3.0.0 and i get the error message that i duplicate the identifiers. Overloaded properties are supported in FPC 2.6.4 and Delphi as well. Is this a bug or a feature in FPC 3.0.0? ---------------------------------------------------------------------- Bitbeisser · 2015-11-27 23:22 Quote from: marcov on November 27, 2015, 08:10:51 pmQuote from: Blaazen on November 27, 2015, 07:59:18 pmThanks. BTW: "Pestering Peacock" is some Ubuntu related joke?  Not really. It was about two years ago during a FPC hackathon when it turned out the hotel where we staying had a bunch of peacocks. They were very noisy and annoying. During that meeting, the plans for 3.0 (and e.g. the unicode implementation of the RTL) were finalized, and sb proposed to call it "Pestering Peacock", and that stuck And I thought you where referring to a taller two legged peacock, who was ruffling his feathers at that hackathon... ---------------------------------------------------------------------- boren · 2015-11-28 04:40 Great news. I wait for FPC 3 for long, now it's available thanks to all developers. Be more famous and famous! Lazarus and Free Pascal! ---------------------------------------------------------------------- HeavyUser · 2015-11-28 04:43 Quote from: Bitbeisser on November 28, 2015, 12:22:09 amAnd I thought you where referring to a taller two legged peacock, who was ruffling his feathers at that hackathon...  erm, why there are beacocks with more legs? Thanks for the release seems one of the most awaited. It looks wonderful. ---------------------------------------------------------------------- jwdietrich · 2015-11-28 07:00 That is very good news.  Thanks for your joint efforts. ---------------------------------------------------------------------- Thaddy · 2015-11-28 08:48 Congrats. I like this approach: "It's ready when it's ready". A breath of fresh air compared to some. The result is a thoroughly modern and stable Pascal environment that duly deserves the 3.0 release number. Very, very good work! Tnx! A special mention should go to the excellent documentation which is way ahead of its siblings. ---------------------------------------------------------------------- gour · 2015-11-28 09:55 Quote from: hnb on November 27, 2015, 09:59:57 amSignificant step forward . Now we need more detailed optional RTTI, anonymous methods and attributes ^^. Bye bye Delphi. Recently I bought Object Pascal Handbook (by Marco Canto) and plan to use it exclusively with FPC/Lazarus, so wonder if there are some more features besides the ones listed above which are not (yet) supported in FPC? ---------------------------------------------------------------------- Luc · 2015-11-28 10:17 This is really great news. Many many thanks to all the FPC team for writing such a good compiler ! ---------------------------------------------------------------------- silvestre · 2015-11-28 11:10 Congratulations. Developers, great job !! ---------------------------------------------------------------------- RNBW · 2015-11-28 11:37 Pestering Peacock looks interesting, but when trying to incorporate into Lazarus, there doesn't appear to be a Sources folder and I keep getting an error message.  I have installed fpc 3.0 to C:\fpc ---------------------------------------------------------------------- marcov · 2015-11-28 12:31 With FPC releases, sources are a separate download. ---------------------------------------------------------------------- Thaddy · 2015-11-28 15:01 Quote from: gour on November 28, 2015, 10:55:03 amQuote from: hnb on November 27, 2015, 09:59:57 amSignificant step forward . Now we need more detailed optional RTTI, anonymous methods and attributes ^^. Bye bye Delphi. Recently I bought Object Pascal Handbook (by Marco Canto) and plan to use it exclusively with FPC/Lazarus, so wonder if there are some more features besides the ones listed above which are not (yet) supported in FPC? Not many. Note that there are and always will be some quirks between FPC and Delphi but not anything (as it stands..) that can't be solved both ways by an experienced programmer that is not afraid to read documentation (ok, there's the limiting factor) or sound advice on this forum. I know these points are being worked on. Check the public SVN source tree. And now for something completely different but on subject and very opinionated: - Anonymous methods are great. Would be very cool to have. Some work done, I believe. - Anything RTTI extra still makes me sick. Use a script language instead, like PHP. - Attributes are even worse for a compiled language. (If you need it, the language is not expressive enough. That is not the case.) Let me explain I use both Delphi and FPC for a very very long time professionally. (Yes, Both) ---------------------------------------------------------------------- gour · 2015-11-28 16:44 Quote from: Thaddy on November 28, 2015, 04:01:07 pmNot many. Note that there are and always will be some quirks between FPC and Delphi but not anything (as it stands..) that can't be solved both ways by an experienced programmer that is not afraid to read documentation (ok, there's the limiting factor) or sound advice on this forum. Good to know. QuoteAnd now for something completely different but on subject and very opinionated: ... Even better. :-) I did not bother to get a closer look what the above mentioned features stand for in Delphi, so it seems FPC is even better in this regard. :-) QuoteLet me explain I use both Delphi and FPC for a very very long time professionally. (Yes, Both) I'm happy to hear it...just curious, what is your development platform for FPC? ---------------------------------------------------------------------- Thaddy · 2015-11-28 16:56 varies between copy con program.pp program XXX;begin end.crtl-z , fp, Notepad, D7, D2007 XE2 and Lazarus. Honest. Very Honest. Depends. ---------------------------------------------------------------------- skalogryz · 2015-11-29 09:37 Quote from: marcov on November 27, 2015, 08:10:51 pmNot really. It was about two years ago during a FPC hackathon when it turned out the hotel where we staying had a bunch of peacocks. They were very noisy and annoying. ... and sb proposed to call it "Pestering Peacock", and that stuck I envision something like "FPC 3.1.0 'Ice Beer' has been released" Personally, I really liked pascal community (fpc, lazarus, mseide ... and so on) staying away from giving fancy names to their releases. I don't really understand this new modern trend of giving meaningless names to releases. ---------------------------------------------------------------------- marcov · 2015-11-29 10:50 Quote from: skalogryz on November 29, 2015, 10:37:15 amI envision something like "FPC 3.1.0 'Ice Beer' has been released" Personally, I really liked pascal community (fpc, lazarus, mseide ... and so on) staying away from giving fancy names to their releases. I don't really understand this new modern trend of giving meaningless names to releases. I guess it was one of a kind, and we go back to pictures of Cheetahs again. ---------------------------------------------------------------------- gour · 2015-11-29 10:52 Quote from: Thaddy on November 28, 2015, 05:56:53 pmvaries between copy con program.pp program XXX;begin end.crtl-z , fp, Notepad, D7, D2007 XE2 and Lazarus. Honest. Very Honest. Depends. So, Windows is main OS, right? ---------------------------------------------------------------------- Thaddy · 2015-11-29 12:40 Nope. It's - server side - Debian based linuxes atm. But my development setup is still windows centric and I use cross-compilation a lot. Also, all test suites are written for both platforms and debugging takes place on the platform itself. ---------------------------------------------------------------------- gour · 2015-11-29 16:09 Quote from: Thaddy on November 29, 2015, 01:40:45 pmNope. It's - server side - Debian based linuxes atm. That's interesting in a sense you have not jumped to the latest called 'Go lang'. ---------------------------------------------------------------------- RNBW · 2015-11-30 20:21 Marcov: Thanks.  I'd not picked up that Sources was a separate download.  All sorted. ---------------------------------------------------------------------- marcov · 2015-11-30 20:34 Quote from: gour on November 29, 2015, 05:09:42 pmThat's interesting in a sense you have not jumped to the latest called 'Go lang'. Maybe he is not writing social networks integrated with search engines :_) ---------------------------------------------------------------------- georgebou · 2015-11-30 21:10 Good job guys. Thank you. ---------------------------------------------------------------------- Thaddy · 2015-12-01 15:14 My only DWIM compilers are pascal based, Marco. Btw what is social? ---------------------------------------------------------------------- marcov · 2015-12-01 16:21 Quote from: Thaddy on December 01, 2015, 04:14:56 pm Btw what is social? A word that is nowadays a higher score on buzzword bingo. ---------------------------------------------------------------------- skalogryz · 2015-12-01 16:59 Quote from: marcov on December 01, 2015, 05:21:55 pmQuote from: Thaddy on December 01, 2015, 04:14:56 pm Btw what is social? A word that is nowadays a higher score on buzzword bingo. Speaking of social! marcov congratulations on 4000th post! ---------------------------------------------------------------------- marcov · 2015-12-01 17:10 Quote from: skalogryz on December 01, 2015, 05:59:35 pmSpeaking of social! marcov congratulations on 4000th post! Thanks  : ---------------------------------------------------------------------- x2nie · 2016-01-12 13:28 Quote from: marcov on November 26, 2015, 08:19:27 pmFPC 3.0.0 ********************************************************************** ****** * Features ********************************************************************** ****** - can call external C code Is it new feature of FPC 3.0.0 ? ---------------------------------------------------------------------- marcov · 2016-01-12 13:46 Quote from: x2nie on January 12, 2016, 02:28:03 pm - can call external C code Is it new feature of FPC 3.0.0 ? No, just like Borland Pascal (and Delphi) compatiiblity a few lines higher isn't new either. IOW that is general description of the project, not whatsnew ---------------------------------------------------------------------- x2nie · 2016-01-12 13:59 Quote from: marcov on January 12, 2016, 02:46:25 pmQuote from: x2nie on January 12, 2016, 02:28:03 pm - can call external C code Is it new feature of FPC 3.0.0 ? No, just like Borland Pascal (and Delphi) compatiiblity a few lines higher isn't new either. IOW that is general description of the project, not whatsnew Oh ! I was wrong, I think it was an implementation of  "future" FPC feature[1], But, it is another one[2]. QuotePlanned for later versionsLinking with C++ code [1] http://www.freepascal.org/future.var [2] ftp://ftp.freepascal.org/fpc/docs-pdf/CinFreePascal.pdf ---------------------------------------------------------------------- BeniBela · 2016-01-27 21:55 Now I tried to update QuoteMakefile:2872: *** The only supported starting compiler version is 3.0.0. You are trying to build with 2.6.2..  Stop. Awesome So you can only can get 3.x, if you already have 3.0 ---------------------------------------------------------------------- Laksen · 2016-01-27 22:34 No, trunk can only be built with 3.0.0. 3.0.0 can be built by 2.6.4. 2.6.4 can be built by 2.6.2 ---------------------------------------------------------------------- Leledumbo · 2016-01-28 10:14 Quote from: BeniBela on January 27, 2016, 10:55:19 pmNow I tried to update QuoteMakefile:2872: *** The only supported starting compiler version is 3.0.0. You are trying to build with 2.6.2..  Stop. Awesome So you can only can get 3.x, if you already have 3.0  The golden rule since a long time ago (but more forced when that Makefile message is introduced) is: previous latest/current stable is the only one supported to build current stable, current stable is the only one supported to build trunk. ---------------------------------------------------------------------- Thaddy · 2016-01-28 10:50 Quote from: BeniBela on January 27, 2016, 10:55:19 pmSo you can only can get 3.x, if you already have 3.0  Yes. And there are many very good reasons for that, most important is: The developers want to make use of the new features in the last major stable release, i.e. 3.0.0 to further develop trunk. e.g.: by now the Default() intrinsic is used in the trunk sourcecode. FPC 2.6.4 (or in your case worse:2.6.2) can't handle that intrinsic because it doesn't know what it is... See the point? So bootstrapping trunk is and always has been enforced to use the latest stable release. Which is as it stands 3.0.0. BTW: the latest fpcup builds you an intermediate 3.0.0 compiler from 2.6.2 to bootstrap trunk. ---------------------------------------------------------------------- marcov · 2016-01-28 11:54 Quote from: Thaddy on January 28, 2016, 11:50:39 am BTW: the latest fpcup builds you an intermediate 3.0.0 compiler from 2.6.2 to bootstrap trunk. Undocumented and unsupported behaviour. The last release before 3.0.0 is 2.6.4 not 2.6.2 ---------------------------------------------------------------------- Thaddy · 2016-01-28 12:10 Yes. It is. But 2.6.4 was a bug fix release and didn't contain new (documented) features. I didn't think it was appropriate to ask to back-port the correct behavior to 3.0.0. You are of course free to back-port it I can only ask... and did ask and submitted a patch  for trunk that got accepted So, basically it was my fault ---------------------------------------------------------------------- molly · 2016-03-01 17:09 I ran into something weird. I have FPC 3.0.0 installed (windows) from official distributed installer, and it came with examples inside examples directory. All the numlib examples seem to have their code duplicated inside the same file. I'm not sure it's something that is amiss with my particular test setup but perhaps someone else could check/verify ? As for an example detgpbte.pas: Code: [Select]program detgpbte; { 1: 260 2: 64 3: } uses   typ,   iom,   det; const   pmin = -10;   pmax = 100; var   l, i, ind, rw, n, k, term, p, vb, nvb: ArbInt;   f: ArbFloat;   a: array[pmin..pmax] of ArbFloat; begin   iom.npos := 1000;        {max. width of output to 1000, since this is piped}   Assign(input, ParamStr(1));   reset(input);   Assign(output, ParamStr(2));   rewrite(output);   Write(' program results detgpbte');   case sizeof(ArbFloat) of     4: writeln('(single)');     6: writeln('(real)');     8: writeln('(double)');     10: writeln('(extended)');   end;   Read(nvb);   writeln;   writeln(' number of examples:', nvb: 3);   for vb := 1 to nvb do   begin     writeln;     writeln('example', vb: 2);     Read(p, n, l);     ind := p;     writeln;     writeln('  n=', n: 1, '  l=', l: 1);     for i := 1 to n do     begin       if i <= l then         rw := i       else         rw := l + 1;       iomrev(input, a[ind], rw);       Inc(ind, rw);     end;     detgpb(n, l, a[p], f, k, term);     ind := p;     writeln;     writeln(' A (left-under) =');     for i := 1 to n do     begin       if i <= l then         rw := i       else         rw := l + 1;       if i > l + 1 then         Write('': (i - l - 1) * (numdig + 2));       iomwrv(output, a[ind], rw, numdig);       Inc(ind, rw);     end;     writeln;     writeln('term=', term: 2);     writeln;     if term = 3 then       writeln(' wrong input')     else     if term = 2 then       writeln(' matrix not pos-def.')     else     begin       Write(' determinant of A =', f: numdig);       {      if k <> 0 then } Write(' * 8**', k: 3);       writeln;     end; {term=1}     writeln('---------------------------------------------');   end; {vb}   Close(input);   Close(output); end. program detgpbte; { 1: 260 2: 64 3: } uses   typ,   iom,   det; const   pmin = -10;   pmax = 100; var   l, i, ind, rw, n, k, term, p, vb, nvb: ArbInt;   f: ArbFloat;   a: array[pmin..pmax] of ArbFloat; begin   iom.npos := 1000;        {max. width of output to 1000, since this is piped}   Assign(input, ParamStr(1));   reset(input);   Assign(output, ParamStr(2));   rewrite(output);   Write(' program results detgpbte');   case sizeof(ArbFloat) of     4: writeln('(single)');     6: writeln('(real)');     8: writeln('(double)');     10: writeln('(extended)');   end;   Read(nvb);   writeln;   writeln(' number of examples:', nvb: 3);   for vb := 1 to nvb do   begin     writeln;     writeln('example', vb: 2);     Read(p, n, l);     ind := p;     writeln;     writeln('  n=', n: 1, '  l=', l: 1);     for i := 1 to n do     begin       if i <= l then         rw := i       else         rw := l + 1;       iomrev(input, a[ind], rw);       Inc(ind, rw);     end;     detgpb(n, l, a[p], f, k, term);     ind := p;     writeln;     writeln(' A (left-under) =');     for i := 1 to n do     begin       if i <= l then         rw := i       else         rw := l + 1;       if i > l + 1 then         Write('': (i - l - 1) * (numdig + 2));       iomwrv(output, a[ind], rw, numdig);       Inc(ind, rw);     end;     writeln;     writeln('term=', term: 2);     writeln;     if term = 3 then       writeln(' wrong input')     else     if term = 2 then       writeln(' matrix not pos-def.')     else     begin       Write(' determinant of A =', f: numdig);       {      if k <> 0 then } Write(' * 8**', k: 3);       writeln;     end; {term=1}     writeln('---------------------------------------------');   end; {vb}   Close(input);   Close(output); end. Thx in advance. ---------------------------------------------------------------------- wp · 2016-03-01 17:18 No fpc3.0 issue, seems to be like this for a long time: my oldest fpc available is 2.0, and it shows the same... ---------------------------------------------------------------------- molly · 2016-03-01 17:22 @wp: thank you very much for the verification. Sorry for the false report. I see. they are wrongly in the svn to begin with :-S ---------------------------------------------------------------------- Semaphore · 2016-03-04 10:14 Thanks and congrats on the 3.0.0 release of FPC, looking forward to using Pestering Peacock We use FPC every day and our team would like to move over to 3.0.0, however we are encountering an issue and I would greatly appreciate confirmation that it's not a problem we can work around on our end. I am referring to http://mantis.freepascal.org/view.php?id=29547 . Any workarounds at all? Cheers guys! ---------------------------------------------------------------------- staratel20 · 2016-03-14 14:07 Thank for your job and forwarding good doing ---------------------------------------------------------------------- Cyrax · 2016-03-14 15:45 Quote from: staratel20 on March 14, 2016, 03:07:56 pmThank for your job and forwarding good doing. Only want to notice a little disappointment, when download a new FPC version and try to install: even if I choose Minimum installation - installer ask me about the path to Android NDK files. Not good if I'm not in plan to write for Android , at least this time. So I have to download Android NDK - 736MB files, that is not simple with my pure internet connection. Also if connection will broken - I'll try again and again until success. Be good to provide download all needed (especially big) things through torrent if possible. And of cause I would be some more glad if Minimum installation of FPC don't need Android NDK (but think if it is - there was no another way). Are you sure you got your FPC installation binary from official source? ---------------------------------------------------------------------- staratel20 · 2016-03-14 15:55 Cyrax: finally clean for me that this not the problem of FPC, so I don't want to interfere to a general joy in this branch)). So propose to remove your last post, and then I remove my. And also want to thank you. ---------------------------------------------------------------------- BeniBela · 2016-03-28 09:14 Quote from: wp on March 01, 2016, 06:18:08 pm my oldest fpc available is 2.0, and it shows the same... I just found 2.0.4 on my hard drive Perhaps now is time to delete that as the hd running out of space ---------------------------------------------------------------------- StamleyMarcus · 2016-05-19 14:40 In this announcement, I see: Free Pascal 3.0.0 is currently available for the following platforms: ... - AIX 5.3 and later for PowerPC (32 and 64 bit) ... But under "Downloads" I cannot find any downloads for AIX ... Can someone please indicate what needs to be done to get this FreePascal compiler to run under AIX? Thanks in advance, Stan ---------------------------------------------------------------------- Thaddy · 2016-05-19 15:46 Normally you would build a cross compiler for AIX from either Linux or Windows or maybe OSX. If your AIX system has sufficient resources you may be able to build a native system using the cross compiler to build the system. ---------------------------------------------------------------------- mischi · 2016-05-19 19:42 The hardest part for me is to build the cross binutils, in particular if it is not linux or freebsd. Once this is done, the cross-compiler is build fairly easily from the sources. ---------------------------------------------------------------------- Paul Breneman · 2016-05-19 20:02 Quote from: StamleyMarcus on May 19, 2016, 04:40:57 pmCan someone please indicate what needs to be done to get this FreePascal compiler to run under AIX? I see 7 powerpc-aix files here: http://svn2.freepascal.org/svn/fpcbuild/binaries/i386-win32/ The easiest way I know to compile FPC for cross-compiling programs for other devices is here: http://wiki.freepascal.org/Small_Virtual_Machines ---------------------------------------------------------------------- mischi · 2016-05-20 05:58 I had a go for the cross-compiler, but it failed with: pp.pas(219,36) Warning: Object /lib/crt0.o not found, Linking may fail ! /usr/local/bin/powerpc-aix-ld: cannot find /lib/crt0.o /usr/local/bin/powerpc-aix-ld: cannot find -lm /usr/local/bin/powerpc-aix-ld: cannot find -lc /usr/local/bin/powerpc-aix-ld: cannot find -lbsd An error occurred while linking The commands to get here are: Building the cross-binutils: ./configure --target=powerpc-aix --prefix=/usr/local/lib/powerpc-aix --bindir=/usr/local/bin --program-prefix=powerpc-aix- --disable-werror make sudo make install Building the cross-compiler: make all OPT="-ap" CPU_TARGET=powerpc OS_TARGET=aix Maybe this helps. ---------------------------------------------------------------------- deathshadow · 2016-07-21 23:22 Please, for the love of Christmas, drop the goofy Ubuntu style cutsy name. Side note, at first I thought it said "Festering Peacock". Could be worse though, could be Abner Audubon Peacock.... ---------------------------------------------------------------------- marcov · 2016-07-22 08:28 Quote from: deathshadow on July 22, 2016, 01:22:26 amPlease, for the love of Christmas, drop the goofy Ubuntu style cutsy name. It is actually a pun on those names. We were discussing if the next major release would be 2.8 or 3.0 at a conference center where was a peacock on grounds making a lot of noise :-) Don't worry, as far as I know it is a one time joke. I haven't heard anyone suggesting a new name for 3.0.2. ---------------------------------------------------------------------- Leledumbo · 2016-07-22 08:42 Quote from: marcov on July 22, 2016, 10:28:08 amQuote from: deathshadow on July 22, 2016, 01:22:26 amPlease, for the love of Christmas, drop the goofy Ubuntu style cutsy name. It is actually a pun on those names. We were discussing if the next major release would be 2.8 or 3.0 at a conference center where was a peacock on grounds making a lot of noise :-) Don't worry, as far as I know it is a one time joke. I haven't heard anyone suggesting a new name for 3.0.2. No problem with Ubuntu style name for me. Continue with 3.0.2, please ---------------------------------------------------------------------- Thaddy · 2016-07-22 08:52 Quote from: Leledumbo on July 22, 2016, 10:42:29 amNo problem with Ubuntu style name for me. Continue with 3.0.2, please I smell a contest..... Submission: "Wandering Willow" ---------------------------------------------------------------------- marcov · 2016-07-22 09:22 Quote from: Thaddy on July 22, 2016, 10:52:09 amI smell a contest..... Submission: "Wandering Willow" Sure, no problem. If next conference a willow walks by we will ---------------------------------------------------------------------- Thaddy · 2016-07-22 10:19 That can be arranged (Halloween costume rental). Or play sims ---------------------------------------------------------------------- marcov · 2016-07-22 12:48 A fake willow gets you a fake release, IOW then we will call rc1 that way. ---------------------------------------------------------------------- Thaddy · 2016-08-06 08:23 I see a 3.0.2 RC1 in svn. Is that already the official RC1? ---------------------------------------------------------------------- marcov · 2016-08-06 19:17 Quote from: Thaddy on August 06, 2016, 10:23:05 amI see a 3.0.2 RC1 in svn. Is that already the official RC1? Probably yes, since sources are already uploaded. Unfortunately, the documentation tools needed some fixes to parse the TArray and advanced records TRect declarations. We are waiting for that (afaik Michael and Mattias are working on that) ---------------------------------------------------------------------- kazalex · 2016-08-06 19:53 I only hope new version will not be released with such bug. ---------------------------------------------------------------------- tom.chris · 2016-08-11 10:16 ImageSize isn't implemented yet in FPC 3.0.0 ---------------------------------------------------------------------- BeniBela · 2016-08-11 11:52 Quote from: kazalex on August 06, 2016, 09:53:41 pmI only hope new version will not be released with such bug. Just do not use nested types Even if fpc would work, gdb won't and then you cannot debug them. ---------------------------------------------------------------------- kazalex · 2016-08-24 07:37 QuoteJust do not use nested typesBug of nested types - just do not use it. Bug of generics - just do not use it. Bug of anything - just do not use it. It's fine. I like it. Hello, Delphi 7! ---------------------------------------------------------------------- Leledumbo · 2016-08-24 08:24 Quote from: kazalex on August 24, 2016, 09:37:54 amQuoteJust do not use nested typesBug of nested types - just do not use it. Bug of generics - just do not use it. Bug of anything - just do not use it. It's fine. I like it. Hello, Delphi 7! Just do not use whatever incomplete features you can't bear of its drawback for now. Feel free to use nested types & generics, despite they choke codetools and the debugger. I use generics, seldomly nested type as well because they still make my code simpler, while I still can use old style debugging when needed. ---------------------------------------------------------------------- SnoopyDog · 2016-10-11 11:25 Unfortunately, FPC 3.0 does not support overloaded properties anymore. Code: [Select]type cTestClass = class   function GetValue (Idx : integer) : string; overload; virtual; abstract;   function GetValue (Key : string) : string; overload; virtual; abstract;   property Value [Idx : integer] : string read GetValue; {$IFNDEF FPC }default;{$ENDIF}   property Value [Key : string] : string read GetValue; {$IFNDEF FPC }default;{$ENDIF} end; This code compiles and works in Delphi 2007, Rad Studio 10.1 and FPC 2.6.4. But not in FPC 3.0 anymore. I wrote it already here: http://forum.lazarus.freepascal.org/index.php/topic,32083.0.html but maybe in this thread it is better (?) ---------------------------------------------------------------------- marcov · 2016-10-11 11:45 No this is not the correct place. For bugs please use the bugtracker. Bugs needs to be fixed in trunk first before we even can talk about the release branch. ---------------------------------------------------------------------- SnoopyDog · 2016-10-11 12:34 Ok thank you! So i need to register at the bugtracker. ---------------------------------------------------------------------- Thaddy · 2016-11-12 19:10 http://bugs.freepascal.org/view.php?id=29056 http://bugs.freepascal.org/view.php?id=30088 Known issue.