What is the most widely used Pascal on Linux and other Unix variants?

Is it Free Pascal?

2012/2/27 Frank Church :

Is it Free Pascal?

Yes, in Linux FPC should have 99%+ of the Pascal market share. Lazarus itself also has at least 95%+ of the IDE/framework marketshare.

For Mac OS X only of all UNIXes the numbers might change if Embarcadero starts making success here, but I am not aware of how well they are doing. But I doubt that they will go very far considering that they don't even have a native IDE, they support only cross-compiling from Windows and only for FireMonkey, not for the VCL.

Op 27 februari 2012 11:33 heeft Felipe Monteiro de Carvalho het volgende geschreven:

Maybe it should, but what makes you think it has 'in fact' such a share?

Vincent

Vincent Snijders wrote

> Yes, in Linux FPC should have 99%+ of the Pascal market share. Lazarus > itself also has at least 95%+ of the IDE/framework marketshare. Maybe it should, but what makes you think it has 'in fact' ?such a share?

Do you know any other Pascal compiler for Linux which competes seriously against FPC? Any big or medium sized project uses that compiler?

Kylix is dead. GNU Pascal is almost dead. So there is not much competition left.

An indirect measure: check http://www.ohloh.net/ All top Pascal developers for non-Windows projects are using FPC, actually most of the top ones are Lazarus developers.

Another indirect measure: All Pascal projects which appeared in the Linux Questions Awards were built in FPC. All but 1 FPC projects there are build using Lazarus+LCL.

Another indirect measure: Put "pascal compiler linux" in Google. I see only mentions of FPC and GNU Pascal (which is almost dead so can hardly count).

I think Vincent is asking where you got the numbers you quoted. Or did you just make them up? I.e. "citation needed".

Henry Vermaak wrote

I think Vincent is asking where you got the numbers you quoted. ?Or did you just make them up? ?I.e. "citation needed".

I invented the numbers. But if anyone thinks they are not realistic, just get the download numbers for FPC and Lazarus, compare with other competing projects, and prove me wrong.

(actually I just took a look and based on what I see I would increase my estimate of the Lazarus share to 99%+ too)

No, that's not how it works. You have to prove it because you stated it. Otherwise I can say stuff like "in the center of the earth lives a dragon named Fred, if you don't think that's correct, you are free to prove me wrong".

NB: I'm not doubting the popularity of fpc/lazarus, but you can't just make up statistics.

Henry

Op 27 februari 2012 12:45 heeft Felipe Monteiro de Carvalho het volgende geschreven:

On debian, it is not so popular compared with gpc: Look at http://qa.debian.org/popcon.php?package=fpc and http://qa.debian.org/popcon.php?package=gpc-4.1

I'd say fpc is about 3-4 (and not 20-100) times as popular based on these figures, so fpc holding maximum 80% of the market share.

Vincent

Huh ? The way I read it, it's exactly opposite ?

Inst number: fp-compiler has 1043, (0,85%), gpc has 353 (0,29%) ?

Votes: fp-compiler has 183, gpc has 47

In each case, it's not a very clear statistic by any standards.

Michael.

Op 27 februari 2012 13:18 heeft het
volgende geschreven:

With 'not *so* popular' I meant having 95 or 99% of the market share.

No, but it is better than nothing.

Vincent

Ehm, market share of what ? Pascal compilers ?
In that case, it's FPC 74 % versus gpc 26 %, which I think is not so bad ?

Of course, total world domination is still far off ;-)

Michael.

Felipe Monteiro de Carvalho wrote

Kylix is dead.

Some products just know know when to die, do they? ;-) I actually know of quite a few commercial applications still written with Kylix today. Beyond Compare 3 from Scooter Software would be an obvious one.

The fact is (as Henry indicated), one can't just imagine a number. Statistics don't work that way.

It's the same as those people that keep on saying that Linux only has 1% market share. How do they know? There are no OEM numbers for all Linux distros. I download my ISO images and share them to many many people. We don't "register" with some company or website to be counted. So that 1% figure is pure speculation! South Africa has many Freedom Toasters, where you bring your own CD's or DVD's and copy any free and open source software at no cost.

Last I read (from a leaked report to Borland and Embarcadero - I can try and find that link again), Borland sold a couple million copies or Kylix (excluding the Open Edition which you could download for free). I'm pretty sure some of those copies are still in use today.

I also host a copy of Kylix 3 Open Edition on some file share website. It was downloaded 70+ times last year, even though I put a notice to say they should try Free Pascal and Lazarus instead.

Henry Vermaak wrote

NB: I'm not doubting the popularity of fpc/lazarus, but you can't just make up statistics.

+1 on both counts.

Op 27 februari 2012 13:30 heeft het
volgende geschreven:

I agree. I disagree, or a better: I seriously doubt what Felipe wrote: "in Linux FPC should have 99%+ of the Pascal market share. Lazarus itself also has at least 95%+ of the IDE/framework marketshare."

Vincent

Take a look at any RPM package search page:

http://rpmfind.net/

Found 157 RPM for fpc

Found 3 RPM for gpc (all of them in very old Mandriva versions)

You can't even find GPC packages in newer distributions.

See also:

http://pkgs.org/search/?keyword=gpc

Appears in Ubuntu 10. Disappeared from Ubuntu 11

Also according to the GNU Pascal site it seems to only support GCC <4 with the last release being from 2006. I can't tell how valid this observations are (I didn't find a source repository by quickly looking), so please handle with care.

Regards,
Sven

What license are FPC and Lazarus, are they GPL?

I think somehow a way must be found of getting Pascal identified with Free Pascal on the Linux platform if it is the most widely used. How about Lazarus Pascal. The problem with Free Pascal, actually most of the older languages is that there tend to be many variations and some what different dialects. There is more or less only one Ruby, one Python, one PHP, one Scala, you know whatever. There are a few variants of Ruby but so long as they can all run Ruby on Rails who cares?

This how I see things.

1. Establish Free Pascal as THE Pascal, THE Object (based) Pascal on Linux

2. A way must be found of uncoupling the Free Pascal, the LCL, the FCL and the Lazarus IDE.

3. The Lazarus IDE (which should be the killer app) must be clearly distinguished. a) as Pascal IDE , ie dealing purely with Pascal Source code, b) an IDE that integrates well with the Non Visual aspects of the LCL e.g. fcl-web for instance c) a graphic based IDE akin to Delphi

4. Both WIKIs are need a makeover. I am sure this has been discussed before :), but the image is really important. Far lesser projects somehow project a more 'professional' image just because of their websites. Lazarus wiki is like the Foyles bookshop of the past if not the present,or like some kind of army surplus store. You can find nearly everything you want, probably everything but it is not as organized and as slick as the competition. A lot of the websites of other projects hardly contain anything, but they all look modern and up to date.

5. This requires an increase in the uptake of Pascal. I mean if a language like D can get so much attention and have libraries being created for it why can't Pascal which has been longer established.

6. I guess one major shortcoming of Pascal is it is not immediately identified with objects, like C. Can Free Pascal simply change its name to Object Pascal
Honestly I think the name is probably the biggest problem if in an era of objects everything it is not associated with Pascal due to its age and past.

In short how does Pascal get itself restablished?

======================= http://devblog.brahmancreations.com

Frank Church wrote

What license are FPC and Lazarus, are they GPL?

The compiler and the IDE themselves are GPL, but their libraries (RTL, FCL, LCL) are released under a modified LGPL (to allow static linking in commercial projects)

2. A way must be found of uncoupling the Free Pascal, the LCL, the FCL and the Lazarus IDE.

FPC and Lazarus are already rather decoupled. They basically only share the Wiki, bug tracker and the SVN server (and some developers ^^).

In that case I prefer "contains everything and does not look modern" ;)

5. This requires an increase in the uptake of Pascal. I mean if a language like D can get so much attention and have libraries being created for it why can't Pascal which has been longer established.

Because
a) Pascal is an old language. It's not stylish and cool to write in an old language except it's C
b) The original Pascal was designed as a lerner language and this is still present in the minds. Thus most people (if they don't know Delphi) think of Pascal as a top language that you can't do any serious programming in

6. I guess one major shortcoming of Pascal is it is not immediately identified with objects, like C. Can Free Pascal simply change its name to Object Pascal

The name "Free Pascal" has already been established as some kind of "Trademark". Also "Object Pascal" is already the name of the dialect FPC supports (and it does not only support Object Pascal, but also the Turbo Pascal and MacPascal dialects). So I personally consider it a really bad idea to change its name.

If a project can not stand out by its name (as you assume for "Free Pascal") it must rely on other things ("features") like stability, performance, portability, etc.

To increase the perceiption of FPC in the public one needs to do advertising. I myself have printed me a FPC T-Shirt and like to run around in that very often (especially at university to oppose the omnipresent Java ;) ).

Regards,
Sven

2012/2/27 Frank Church :

1. Establish Free Pascal as THE Pascal, THE Object (based) Pascal on Linux

While I disagree that any serious competition to FPC exists in Linux, you should know that FPC has never been monopolist. Competition is widely seen as something good. Having more good Object Pascal compilers would be good.

2. A way must be found of uncoupling the Free Pascal, the LCL, the FCL and the Lazarus IDE.

So you are suggesting to make it harder to configure as a step towards higher popularity?

4. Both WIKIs are need a makeover.

There is only 1 wiki and I think it is quite good. All necessary information is quickly accessible. A beginner can very quickly get a quick start on almost any topic.

In short how does Pascal get itself restablished?

I think that everything possible in the world about this topic has already been said in this thread:

http://www.lazarus.freepascal.org/index.php/topic,13754.0.html

Frank Church wrote

dialects. There is more or less only one Ruby,

And which one is that, JRuby(*), IronRuby(**), or Ruby-on-parrot(***))?

(*) http://www.jruby.org Reworked On top of Java
(*) http://ironruby.net/ reworekd on top of .NET
(**) http://cardinal2.rubyforge.org/ reworked on top of Parrot VM

( :-) )

If the rumours are correct, Embarcadero might enter the Linux market again in august-september, but again a crosscompile product.

At least that was how the roadmap during the XE2 introduction sounded, next edition native ARM (iOS and maybe Android) and expansion of the i386/x86_64 compiler to Linux.

While GPC is not very active atm, and a case could be made that it is dead, these kinds of statistics are not very comparable.

This because GPC has a totally different release and development model (read: none since 2.1), and nearly all builds are independent build of target-maintainers, that often also keep own source trees, and mutually absorb patches.

It is very difficult to compare this with FPC, that actively pushes distribution specific releases (deb, rpm, freebsd ports).

I'm not sure if products that hinge on the last version of a series are really relevant.

I'm sure we'll still find many VB6 products in 2025 :-)

Marco van de Voort wrote

I'm not sure if products that hinge on the last version of a series are really relevant.

My point was more about the fact that Scooter Software has applied there own fixes to CLX, and have also upgraded their Kylix 3 to use Qt3 (instead of Qt 2.3 which came with Kylix). If they shared all there Kylix enhancements, I have no idea. I also know there is a huge bunch of "unofficial fixes" that can be installed over Kylix 3. So at least the framework around Kylix has grown, after is was "considered" to be dead.

Also, Kylix 3 is still perfectly fine for web development work.

Graeme Geldenhuys wrote

My point was more about the fact that Scooter Software has applied there own fixes to CLX, and have also upgraded their Kylix 3 to use Qt3 (instead of Qt 2.3 which came with Kylix).

Actually, the upgrade to Qt 3 was Andreas Hausladen's doing, and he did release it, but it isn't available online anymore. It was an impressive piece of work, and I think it laid the groundwork for Lazarus' Qt support, since I remember seeing Zeljan's name associated with it too.

We do have a shocking number of fixes on top of it though.

If they shared all there Kylix enhancements, I have no idea.

No, we haven't. Borland screwed up the Kylix community project, and Andreas stopped maintaining the unofficial patches at the same time, so there isn't anywhere for us to share them.

So at least the framework around Kylix has grown, after is was "considered" to be dead.

It was "considered" dead, and then the Kylix community project and Borland's poor handling of Simon Kissel's cross-compiler killed it completely.

Also, Kylix 3 is still perfectly fine for web development work.

I'm not aware of any Linux distro post-2006 that runs the IDE, and even then you needed a hack to debug threads.

We would have given it up years ago, but I thought Delphi's cross-platform support would use Qt right up until they purchased VGScene. I'm sure there are a few other holdouts, but there's certainly no community anymore. Before they removed it, the .kylix newsgroup's last post was in 2008 or so.

Dear colleagues:

i have used por many years the upx for compact the executable files. Really, it have worked pretty well.

today, i have tried to compact an executable but i have obtained a nom good result.

previously i have used the strip to reduce the "weight" from more or less 30 Mb to 3 Mb. After that, the executable still works well.

them i have applied the upx, it look to work well, the weight is reduced to 1 Mb, but when i tried to execute, a message appears telling "the configuration is not good and maybe reinstalling the application can work". In more than 20 years using it, it is the first time this occurs.

thanks very much

Ing. H?ctor F. Fiandor Rosario

While there are updates for K3 + Qt3 support (even I contributed and used that) there's another problem: you cannot run K3 on any distro with glibc >= 2.4, and that's problem (so it must be at least 6-7 yrs old distro).

zeljko
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120228/638bdd3c/attachment.html>

Yes, I was involved in Qt3 support for Kylix, but it have nothing to do with lazarus qtlcl , since Qt3 used different (and better approach IMO) to generate C bindings. You get all clasees (even protected methods), so overriding was piece of cake - and yes K3/Qt3 apps were snappy and stable :)

Afair, at that time Simon, Andreas and me asked Borland for and NDA .. I was ready to sit down and fix Kylix for free - it costs Borland exactly $0, but they answered after few months that Kylix is not their priority (and that means = dead)....all happened after M$ was involved into borland ... after that .Net delphi was out etc etc ....
So Kylix is dead because of political reasons - nothing else.

zeljko

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120228/e452e4de/attachment-0001.html>

zeljko wrote

While there are updates for K3 + Qt3 support (even I contributed and used that) there's another problem: you cannot run K3 on any distro with glibc >= 2.4, and that's problem (so it must be at least 6-7 yrs old distro).

I don't really see that as a issue only relates
to the Kylix 3 IDE, not the applications you build with it. I can still run [and do do so] Kylix 3 built applications on my Ubuntu 10.04 64-bit system. Simply use a VM for Kylix 3 development. I currently have a Red Hat 9 VM with Kylix 3 Ent setup for exactly that reason. Hell, Embarcadero is even recommending that for Windows and Mac OS X development! I know of many Delphi developers that run Linux (or Mac OS X), but use a Windows VM to run Delphi - simply to get the stability and virus free environment of Linux when they surf the web or check emails. It also keeps your development environment in a pristine condition.

I'm not saying Kylix 3 is still a booming product (it isn't), I'm just saying that some companies invested heavily in Kylix or CLX development, and simply can't move away from it.

Graeme Geldenhuys wrote

I'm not saying Kylix 3 is still a booming product (it isn't), I'm just saying that some companies invested heavily in Kylix or CLX development, and simply can't move away from it.

My company also invested a lot into kylix, but successfully moved to lazarus qtlcl :)

zeljko
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120228/f970bea2/attachment.html>

Marco van de Voort wrote

While GPC is not very active atm, and a case could be made that it is dead,

They (unfortunately) did not follow FPC on the path to Object Pascal (Delphi and Apple style). If FPC programs could be compiled using gcc, we could port Lazarus / FPC programs to any CPU architecture.

-Michael

Hector,

Buenos dias.

Yo uso en Opciones de proyecto/Enlazando marco casilla usar archivo externo de simbolos para reducir el peso del ejecutable.

[I use Proyect Option/Liniking click on Use external dgb debug symbol file to reduce the executable wheight]

Lo habias probado?

Un saludo,

On Tue, Feb 28, 2012 at 1:31 AM, Ing. H?ctor F. Fiandor Rosario <

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120228/9d6b129c/attachment.html>

That's why I wrote "please handle with care" ;)

Regards,
Sven

As recently the use of UPX for packing the FPC binaries was removed I'd suggest you to not rely on it anymore.

Regards,
Sven

PS: Please don't hijack other threads.

My company also tried (and abandoned) Kylix:
http://port2laz.blogspot.com/2005/12/from-clx-to-vcl.html

2012/2/28 zeljko

Same here.
Kylix had a great debugger, something that still doesnt work nice in Lazarus, but other than that Lazarus is exceptional and the support is very good. Try to get any support out of Borland and their successors. Even earlier years before Delphi, they failed to fix RT-200 for pentium and above processors. Easy fix that will take a few minutes, but they rather decided to ignore all their customers.

Lazarus and Free Pascal is just dandy.

Alexsander Rosa wrote

My company also tried (and abandoned) Kylix: http://port2laz.blogspot.com/2005/12/from-clx-to-vcl.html

Michael Schnell wrote

> dead, They (unfortunately) did not follow FPC on the path to Object Pascal (Delphi and Apple style). If FPC programs could be compiled using gcc, we could port Lazarus / FPC programs to any CPU architecture.

More likely, we would, like GPC, have a worthless windows port, and constantly fixing the breakage introduced by new gcc versions.

2012/2/28 Ing. H?ctor F. Fiandor Rosario :

when i tried to execute, a message appears telling "the configuration is not good and maybe reinstalling the application can work".

I don't know what the concrete problem with upx is but the error message really means that windows cannot find one of the dlls referenced in one of the the manifest files or the manifests of referenced dlls, it can happen even if all dlls are installed where they belong (or where you think they might belong) and only some minor version number mismatch or some wrong paths or file names, it's one of the symptoms of MS DLL-Hell V2.0 and is following the well known Microsoft tradition of clear and precise and meaningful error messages.

[OT]
: Let's see these messages... oh dear, so many and so little time. [scrolls document]
What does this one mean?
Engineer : 'Could not find %s in %s in %s' for this context can mean that we cannot locate the function to call in library located in directory where we were looking for it. Manager : [oh dear so little time] but what does this mean ? Engineer : means we found library but it did not contain what we wanted or ... Manager : [ oh dear oh dear so little time] ok but why is that ? Engineer : [stressed] oh... possibly configuration problem ? Manager : oh. ok make it say so then. Gotta go. Make it less cryptic for laymen. Engineer : [facepalm].

[Some long long time way before that]
[Somewhere else in foreign branch]
Manager : so all translations done?
Translator : yeah but this one I don't understand... Manager : we'll let engineers look at it, now close it we gotta ship...

Result : polish translation of 'Not enough quota to continue' translated literally as 'Not enough quotation marks to continue' (translated from 'quota' <=> multiple of 'quote' = cudzyslow = quotation mark) message somebody spotted in NT4...

[/OT]

A famous one in brazilian portuguese:
http://img205.imageshack.us/img205/351/win2000errorph2.jpg

Probably something like "Can't find <some shared folder>. Check the spelling and try again, etc" in the original. Spelling = Ortografia. The message uses the word "orografia", without the T, that exists but with another meaning, like "ortography" vs "orography". It's funny that the computer accuses you of mispelling something but the message itself has a spelling error.

2012/3/7 Lukasz Sokol