Anders Melander 1782 Posted April 6, 2019 It's not possible to downvote in JIRA. 2 Share this post Link to post
David Heffernan 2345 Posted April 6, 2019 Pretty bad reasoning. If it's for paid customers only, then there's no security in case Emba go bust. "Because we love Delphi more than you do." How to win friends and influence people. 4 Share this post Link to post
AlekXL 8 Posted April 6, 2019 (edited) 2 hours ago, David Heffernan said: Pretty bad reasoning. If it's for paid customers only, then there's no security in case Emba go bust. for paid customers there will be some security. They would be able to continue in-house development. Any every long road starts with first step. 2 hours ago, David Heffernan said: "Because we love Delphi more than you do." How to win friends and influence people. Win friends among whom? You should learn to think, not just read. I am winning friends with this very sentence, among those who can vote. And votes are essential in this case Edited April 6, 2019 by AlekXL Share this post Link to post
Bill Meyer 337 Posted April 7, 2019 2 hours ago, AlekXL said: Win friends among whom? You should learn to think, not just read. I am winning friends with this very sentence, among those who can vote. And votes are essential in this case While you antagonize those who will make the actual decision? Votes can be ignored, and have been for years. 3 Share this post Link to post
AlekXL 8 Posted April 7, 2019 such decision is beyond marco let alone coders. its up to suits. and my sentence wont antagonize them - they dont love delphi, and would admit it easily. it's about reaching that level of decision making marco was against inline variables… but there was issue with votes, and i even wrote a page in russian wikipedia with criticism about lacking inline and having arc Share this post Link to post
Fritzew 51 Posted April 7, 2019 There is FPC, spend your time to help there. Then you will have your compiler 4 Share this post Link to post
AlekXL 8 Posted April 7, 2019 FPC has some issue with it's developers. See, some guys contribute features like Anonymous methods or Attributes, but their work silently rejected for no visible reason https://bugs.freepascal.org/view.php?id=24481 https://bugs.freepascal.org/view.php?id=33384 And I cannot do everything on my own.. Moreover, FPC is backed by GDB, which is outright terrible. So what, should I commit myself to merging everything into trunk, and then create some bindings with, say, LLDB -- all on my own? This is fool's errand Share this post Link to post
dummzeuch 1505 Posted April 7, 2019 1 hour ago, AlekXL said: This is fool's errand So is trying to get the Delphi compiler open sourced. 6 Share this post Link to post
AlekXL 8 Posted April 8, 2019 (edited) 10 hours ago, dummzeuch said: So is trying to get the Delphi compiler open sourced. One must try, that's what I believe. If my proposal is to be rejected, so be it, I just need to do my utmost for this or either to happen, and then consider other options. At least I wouldn't despise myself for never even trying, just others, you know. Edited April 8, 2019 by AlekXL Share this post Link to post
Larry Hengen 39 Posted April 8, 2019 Open sourcing the compiler(s) will never happen. That is the core of EMBT's intellectual property, and one of the major reasons why Lazarus/FPC has not eroded the Delphi market further than it has; compiler compatibility. There is no money in writing an IDE and there are lots of them freely available. EMBT's revenue is from developers who have lots of source code which is tied to the Delphi compilers. The IDE makes for most of the development experience, so I think you would have more chance of getting the IDE Core open sourced (like Visual Studio shell) than getting the compiler open sourced. Many developers could competently work on the IDE to improve it, make it portable (xPlatform) etc, but not many developers could contribute to the compilers, IMHO. 2 Share this post Link to post
Joseph MItzen 251 Posted April 8, 2019 On 4/6/2019 at 10:00 PM, AlekXL said: marco was against inline variables… Really? He was praising them in his blog post. 1 Share this post Link to post
Joseph MItzen 251 Posted April 8, 2019 38 minutes ago, Larry Hengen said: That is the core of EMBT's intellectual property, and one of the major reasons why Lazarus/FPC has not eroded the Delphi market further than it has; compiler compatibility. There is no money in writing an IDE and there are lots of them freely available. Isn't the core really the framework? There are several compilers available for dialects of Pascal (Oxygene, FPC, etc.) Even if FPC was a 100% clone of the current Delphi language it wouldn't be much use to legacy customers without a compatible VCL implementation to go along with it. Regards money in writing an IDE, JetBrains is doing just fine selling IDEs... so many, in fact, that they can charge individuals 1/15 the cost of Delphi Pro and offer open source versions of their Python and Java IDEs and even sell a subscription to all of their compilers for far less than Delphi costs. In fact, they can also open source their Kotlin language under the reasoning that they'll be the best source for an IDE for it. And they'll also give away 10-user versions of their bug tracking and project management software, yet somehow they haven't raised prices in at least three years while Delphi's renewal cost goes up 4% every year. That said, the bulk of Embarcadero's Delphi customers are long-time users locked into a large amount of legacy code. There's no real financial benefit to EMBT to open sourcing the compiler. It would be of no benefit to the end users either. Former EMBT employee Barry Kelly inadvertently revealed that the 32bit compiler is over one million lines of undocumented (!!!) ANSI C code. Good luck trying to form a community around that! If they wanted to help they could produce a formal specification for the language, something they've rejected in the past. Then an open source community could leverage LLVM and have a modern, clean compiler produced long before they'd managed to gain any insight into the ancient desktop Delphi compiler's code base. Share this post Link to post
AlekXL 8 Posted April 8, 2019 2 hours ago, Larry Hengen said: Open sourcing the compiler(s) will never happen. That is the core of EMBT's intellectual property... And opensourcing the compiler does not mean EMBT loses it's exclusive copyright. Learn to read Share this post Link to post
Bill Meyer 337 Posted April 8, 2019 15 minutes ago, AlekXL said: And opensourcing the compiler does not mean EMBT loses it's exclusive copyright. Learn to read There you go again, charming those you wish to persuade. As a matter of fact, however, your desire that the compiler be open sourced in no way justifies your assertion that it needs to be open sourced. Share this post Link to post
AlekXL 8 Posted April 8, 2019 (edited) 1 hour ago, Bill Meyer said: There you go again, charming those you wish to persuade. As a matter of fact, however, your desire that the compiler be open sourced in no way justifies your assertion that it needs to be open sourced. but this justifies the need: since 1/1/2018 more 200 open bugs in bugtracker, and only ~30 fixed during that period. Same situation with Feature Requests, actually worse -- apparently they are low-priority. They are drowning. As those who are dependent on their product Edited April 8, 2019 by AlekXL Share this post Link to post
AlekXL 8 Posted April 8, 2019 (edited) Also behold, this priceless, trying to declare bug a feature Marco is really desperate.. Edited April 8, 2019 by AlekXL Share this post Link to post
AlekXL 8 Posted April 8, 2019 3 hours ago, Joseph MItzen said: Former EMBT employee Barry Kelly inadvertently revealed that the 32bit compiler is over one million lines of undocumented (!!!) ANSI C code. Good luck trying to form a community around that! Yeah, I heard that too, but you are underestimating how hardcore some Delphi developer are Share this post Link to post
Joseph MItzen 251 Posted April 8, 2019 1 minute ago, AlekXL said: Yeah, I heard that too, but you are underestimating how hardcore some Delphi developer are But the most hardcore are the ones least willing to admit there are any quality problems. I still believe the best use of resources would be to support NewPascal.... http://newpascal.org/ ...which is based on FPC with a collection of patches the FPC folks won't or haven't gotten around to accepting. It's not easy to get yourself out from vendor lock-in, especially for Delphi users who may have spent 20+ years getting themselves further and further dug in. I know I spent two years disentangling my personal desktop from Windows before I was able to successfully move to Linux. Share this post Link to post
AlekXL 8 Posted April 8, 2019 i considered newpascal, but isnt it effectively dead? the repository has no new commits since summer last year Share this post Link to post
Markus Kinzler 174 Posted April 9, 2019 9 hours ago, AlekXL said: Yeah, I heard that too, but you are underestimating how hardcore some Delphi developer are But this was the "old-gen" compiler beeing subsituted for most of the targets now. Only the Win32 and OS32 compilers are based on it. Share this post Link to post
Alexander Elagin 143 Posted April 9, 2019 I think it would be easier to write a brand new frontend parser from scratch and feed its output to LLVM or whatever else backend code generator because I understand what it means to have to deal with huge amounts of undocumented legacy code. Unfortunately, I highly doubt there are enough developers interested in writing such parser, let alone the absence of the formal grammar definition. And given the strange state of the FPC development where it seems that the core developers are not interested in implementing new features (closures / attributes / etc, even when there are already provided implementations for these features collecting dust in the bugtracker for years) I feel there is nobody to write another Pascal compiler. The RemObjects compiler genius Carlo Kok has already developed Oxygene and I highly doubt anybody could persuade him to develop a free fully Delphi-compatible compiler in his spare time. Share this post Link to post
Dalija Prasnikar 1396 Posted April 9, 2019 23 minutes ago, Markus Kinzler said: But this was the "old-gen" compiler beeing subsituted for most of the targets now. Only the Win32 and OS32 compilers are based on it. No. Only iOS and Android are LLVM based. http://docwiki.embarcadero.com/RADStudio/Rio/en/Delphi_Toolchains AFAIK first Linux release (Tokyo) was also LLVM. And you don't want to have LLVM on other platforms for few reasons. Not only it is slower than "old" handcrafted Delphi compiler, more importantly LLVM cannot handle exceptions in way Delphi compiler does. Using LLVM on desktop would break a whole a lot of code. You can fin more at https://dalijap.blogspot.com/2018/10/catch-me-if-you-can.html Share this post Link to post
David Heffernan 2345 Posted April 9, 2019 1 hour ago, Markus Kinzler said: But this was the "old-gen" compiler beeing subsituted for most of the targets now. Only the Win32 and OS32 compilers are based on it. Win64 also I believe Share this post Link to post
David Heffernan 2345 Posted April 9, 2019 51 minutes ago, Dalija Prasnikar said: No. Only iOS and Android are LLVM based. http://docwiki.embarcadero.com/RADStudio/Rio/en/Delphi_Toolchains AFAIK first Linux release (Tokyo) was also LLVM. And you don't want to have LLVM on other platforms for few reasons. Not only it is slower than "old" handcrafted Delphi compiler, more importantly LLVM cannot handle exceptions in way Delphi compiler does. Using LLVM on desktop would break a whole a lot of code. You can fin more at https://dalijap.blogspot.com/2018/10/catch-me-if-you-can.html That's surely not a problem with LLVM per se, rather the Delphi compiler on top of LLVM? Share this post Link to post