[Issue 15485] switch with no case compiles
https://issues.dlang.org/show_bug.cgi?id=15485 Iain Buclawchanged: What|Removed |Added CC||ibuc...@gdcproject.org Resolution|INVALID |FIXED --- Comment #3 from Iain Buclaw --- Don't even need to bother raising a bug in gdc (by the way this is fixed, not invalid). --
Re: [Semi-OT] I don't want to leave this language!
On Wednesday, 7 December 2016 at 02:38:50 UTC, bpr wrote: It's a counterfactual at this point, but I would guess that if D had left out the GC in 2010 when D2 came out it would have been ahead of C++ in many ways and perhaps would have been able to peel off more C++ programmers c++ programmers want c++. anything that is not c++ will be bashed to death. there is absolutely no reason to kill one of the key D features only to attract 2.5 c++ coders. actually, we already have That One C++ Programmer We Need onboard -- Andrei. ;-)
Re: Should we warn if we detect null derefernces or void value uses ?
On Wednesday, 7 December 2016 at 06:47:56 UTC, burjui wrote: Yes, but we should issue an error, like Andrei said. A language is only as useful as it's best implementation, so diagnostics like that are essential. Data-flow analysis shouldn't even be optional or lacking in any modern compiler. you know, D has variable initialization with default values. that allows to skip DFA, and still be sure that variable is initialized. also, that thing can be done in *backend*, if it is really necessary, along with eliminating needless loads. such warning/error absolutely unnecessary in frontend, it only adds more code to maintain, and slows down compiling.
[Issue 16047] Range violation in setting multi-dimensional AA entries
https://issues.dlang.org/show_bug.cgi?id=16047 MichaelZchanged: What|Removed |Added CC||dlang@bregalad.de --
Re: Should we warn if we detect null derefernces or void value uses ?
On Monday, 5 December 2016 at 04:41:55 UTC, Stefan Koch wrote: Hi Guys, What is your opinion, should we warn if we unambiguously detect something that is clearly unwanted ? int fn(int y) { int x = void; ++x; return x+y; } This requires data-flow analysis (The same kind that tells you if you are skipping a statement) And will slow down compilation a little if we enable such a warning. Yes, but we should issue an error, like Andrei said. A language is only as useful as it's best implementation, so diagnostics like that are essential. Data-flow analysis shouldn't even be optional or lacking in any modern compiler.
Re: How about a bounty for a new windows installer using inno setup ?
On Tuesday, 6 December 2016 at 17:28:25 UTC, Jim Hewes wrote: On 12/6/2016 12:21 AM, Thomas Mader wrote: You can also create a WiX installer out of an InnoSetup installer. I think it's more important to decide upon the feature set, readability and the time needed to build an installer. Have you experience with both? I only have experience with NSIS and InnoSetup and in InnoSetup the feature set for Windows is really good and the readability is good. I started out by using InstallShield some years ago and got battle scars there. I don't recommend that. I used NSIS a little because a company we partnered with required it but I'm no authority on NSIS or Inno. It really depends on how complicated your particular install is and where you expect it to go in the future. If you're just copying a few files then anything will work. I don't mean to make too big a deal out of it if the requirements are really simple. Personally I think it's better in the long run to generate an MSI for several reasons you can probably look up yourself---security, ability to rollback (installation is a transaction), appears in Programs and Features, transforms, etc. I wouldn't advise doing the coding part externally in D this makes things much more complicated than it should be. Stick with what's supported by the tool. I'm not suggesting you necessarily use D together with something like NSIS. But you do want to have a one-button automated build process, not just for convenience but for repeatability. That's important. Soon you will want to get away from the tool's own GUI and run things programmatically. After having not-so-good experiences with InstallShield I looked at things like SCons and msbuild, which was just coming out at the time. (This was a while ago). I tried msbuild but it didn't have modules to support many of the things I needed to do. Things like code-signing with a verisign signature, injecting data and files into exe resources, etc. Fortunately you can build your own custom modules using C# which is what I tried. But the process of transferring variables back and forth from the script to C# for every custom module was painful and I thought, "if I just do this all in C# it will be much easier". So I switched. I used C# not only to call on the WiX tools to run them, but to easily manipulate pathname and filename strings, which were different because I needed to build different configurations for different customer companies. And I also needed to build different combinations of language localization. I could use .NET to build a nice GUI for selecting configurations, and C# to call the Windows API when needed, move files around, anything. You may also run into issues when you need to do complicated updates and there are already earlier versions in the field. You may want to remove features, but your installer has to both update existing users in addition to supporting new users. It's hard to predict the future though so I won't say much about it except that it helps to have a more powerful tool when you run into such situations. The scripting-type tools are tempting because they're easy and no one wants to spend any time on installers. It's usually something that people hope to just slap on at the end and it often gets underestimated. But as I said, maybe it IS easy if you're just copying files and you will only ever have one configuration. So it depends. Jim Nice writeup. In our company we used NSIS and are switching over to InnoSetup. Most of the work is already done, just a few apps need to be moved. I think no one really wants NSIS because of readability issues. It's assembler like language is too low level and many common functions are just missing. InnoSetup on the other hand gives you everything you could ever want. I never missed a thing because pretty much everything is right there and if you happen to have very special needs you can do it quite nicely with the pascal scripting ability. You can call every Windows API function you just need to wrap it if it is not provided in one way or another via the InnoSetup API. I don't think you need to do that for the D Installer though. Our installers need to handle quite a few things. - Signing and timestamping of exes, dlls, Installer and Uninstaller - SendTo Handler registration (Windows doesn't provide a common SendTo directory so you need to handle it quite complicated for each users individual SendTo directory if you don't want to write a proper SendTo COM thing) - Registration of a COM server - Differentiate between 32 and 64 bit installs - User elevation for Installs and Uninstalls but Updates are done for the current User (We still use Inno for our Updates too but move away from that because it's not at all transactional and cannot be undone in the middle of the update) - Create MSI Installer from InnoSetup Installer via Wix The update case could be better supported
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis wrote: So, while there are certainly folks who would prefer using D as a better C without druntime or Phobos, I think that you're seriously overestimating how many folks would be interested in that. Certainly, all of the C++ programmers that I've worked with professionally would have _zero_ interest in D as a better C. - Jonathan M Davis Considering scientific/numerical applications, I do agree with Ilya: it is mandatory to have zero overhead and a straightforward/direct interoperability with C. I am impressed by the Mir lib results and I think "BetterC" is very attractive/important. -- Vincent
Re: CTFE Status
On Tuesday, 6 December 2016 at 23:55:06 UTC, Nordlöw wrote: On Tuesday, 6 December 2016 at 23:16:16 UTC, Stefan Koch wrote: Yes there is. Create a char[] give it the length of string a + string b; copy string a into the char[] then copy string b into the resulting array. offsetting it by the length of string a. Sounds like you have your concat implementation right there ;) not really, concat needs to be smart about allocation and memory reuse. when used on char arrays the memory access will be different. and use more instructions.
Re: DMD 2.072.1 MSVC build seems to be failing
On 12/6/2016 7:38 PM, Lewis wrote: On Wednesday, 7 December 2016 at 02:49:14 UTC, Walter Bright wrote: Those files were converted to glue.d, etc., but obviously the VS build was not updated. Please file a bugzilla issue for it. Done, thanks! https://issues.dlang.org/show_bug.cgi?id=16954 Good! Next, you can also file a PR to fix it! Should be pretty simple.
[Issue 16949] [Reg 2.073] confusing @safe error message for fields with unsafe destructors
https://issues.dlang.org/show_bug.cgi?id=16949 Walter Brightchanged: What|Removed |Added Status|NEW |RESOLVED CC||bugzi...@digitalmars.com Resolution|--- |WORKSFORME --- Comment #1 from Walter Bright --- Already fixed: C:\cbx\bug>..\dmd test7 -transition=safe test7.d(6): Error: @safe destructor 'test7.B.~this' cannot call @system destructor 'test7.A.~this' C:\cbx\bug>..\dmd test7 test7.d(6): Error: @safe destructor 'test7.B.~this' cannot call @system destructor 'test7.A.~this' --
[Issue 13314] BinaryHeap assumes Store has dup property
https://issues.dlang.org/show_bug.cgi?id=13314 --- Comment #2 from github-bugzi...@puremagic.com --- Commits pushed to master at https://github.com/dlang/phobos https://github.com/dlang/phobos/commit/fba056a8df961138859561a6f4ea3d9447efbd6d fix issue 13314 https://github.com/dlang/phobos/commit/9eb35e11b39ed54fea8b6f4fc18503c86bddc05a Merge pull request #4929 from somzzz/issue_13314 fix issue 13314 - BinaryHeap assumes Store has dup property --
Re: DMD 2.072.1 MSVC build seems to be failing
On Wednesday, 7 December 2016 at 02:49:14 UTC, Walter Bright wrote: Those files were converted to glue.d, etc., but obviously the VS build was not updated. Please file a bugzilla issue for it. Done, thanks! https://issues.dlang.org/show_bug.cgi?id=16954
[Issue 16954] New: MSVC build fails on DMD 2.072.1 release
https://issues.dlang.org/show_bug.cgi?id=16954 Issue ID: 16954 Summary: MSVC build fails on DMD 2.072.1 release Product: D Version: D2 Hardware: All OS: Windows Status: NEW Severity: normal Priority: P1 Component: dmd Assignee: nob...@puremagic.com Reporter: musicalje...@gmail.com I downloaded and installed a fresh copy of DMD 2.072.1, opened dmd2/dmd/vcbuild/dmd.sln, and tried to build. I got several compilation errors related to missing files, in particular: 1>c1xx : fatal error C1083: Cannot open source file: '..\glue.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\msc.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\toir.c': No such file or directory According to Walter in https://forum.dlang.org/post/odrwelyfjvhysggee...@forum.dlang.org some files in the build have been translated to .d files, but the VS build probably hasn't been updated accordingly. --
Re: [Semi-OT] I don't want to leave this language!
On Wednesday, December 07, 2016 02:38:50 bpr via Digitalmars-d-learn wrote: > On Tuesday, 6 December 2016 at 22:47:34 UTC, Jonathan M Davis > > We get plenty of folks who aren't big C/C++ programmers who are > > interested in D. Yes, the majority seem to have a C++ > > background, but we also get folks from C#, python, ruby, etc. > > It would be nice to see a breakdown. From where I sit, it appears > that most of the interest in D is from C++ users, and it doesn't > appear that D popularity is rising so much. Any data that belies > that sad assessment? As I understand it, the downloads have been increasing year over year. At least some years, Andrei has given some statistics on the state of D in his dconf talk. And per his stats, forum/newsgroup activity was going up too, though I get the impression that at least in the main newsgroup, it may have actually dropped off (whereas Learn seems like it's been growing). That's just my impression though. I'm sure github activity has gone up over time though. It's my understanding that D's usage has continued to grow but that it isn't growing super fast, whereas languages such as Go and Rust do seem to have grown very quickly (probably at least in part due to the companies behind them). But other languages have taken a long time to catch on and still ended up being very successful with large user bases (e.g. that's what happened with python). It's also hard to gauge how much D is really being used. The number of companies saying that they're using D seems to have increased, but there are who-knows how many folks using D who haven't said anything, and we really don't have much to go on besides the downloads, which only capture a portion of the D's users and even then only tells you how often dmd was downloaded, not how many people it was or if they're new users or whatnot. And downloading dmd from dlang.org is not the only way to get it. There are definitely things that we can be and should be doing to improve D's traction (like better supporting @nogc in Phobos), but I don't think that we're doing badly. And often, the problem seems to be more of a PR one than anything technical (e.g. I think that we're finally pretty much beyond the issues caused by the confusion over Tango vs Phobos in D1, but it took a long time). And honestly, much as there are technical problems related to the GC, the far bigger problem seems to be the PR issues related to it. Sadly, the simple fact that we _have_ a GC has been a PR problem, regardless of the actual state of things. But regardless, we do seem to be gaining traction, even if it's not as quickly as we might like. - Jonathan M Davis
Re: [Semi-OT] I don't want to leave this language!
On Wednesday, 7 December 2016 at 02:24:56 UTC, bpr wrote: If I really *want* to use a GC, say I'm writing a server and I believe that a well tuned GC will allow my server to stay alive much longer with less fragmentation, I'll probably skip D and pick Go or maybe (hmmm...) even Java because their GCs have had a lot of engineering effort. Writing a server is quite narrow compared to the "programmers who are comfortable working in a GC-ed language" that I was responding to. I wonder what percentage of Ruby programmers have thought about garbage collection ever. Why would a Ruby or Python programmer unconcerned with performance want to switch to D? I'm sure there are some who would, but I'd imagine they're rare. Maybe some prefer D as a language? The same argument could be used against any language. Performance is far from the only reason to use D.
Re: DMD 2.072.1 MSVC build seems to be failing
On 12/6/2016 6:15 PM, Lewis wrote: I like to build DMD with MSVC because it complies a little faster than the shipped executable. I downloaded and installed a fresh copy of DMD 2.072.1, opened dmd2/dmd/vcbuild/dmd.sln, and tried to build. I got several compilation errors related to missing files, in particular: 1>c1xx : fatal error C1083: Cannot open source file: '..\glue.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\msc.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\toir.c': No such file or directory I tried fiddling a bit with removing these files from the project, and I could get the first stage of the build to pass, but then I ran into linker errors later on. I don't know enough about how this project is set up, so I figured I'd just ask. Any ideas what's going on? Did the VS project simply not get updated for the new release, or am I missing something in my build procedure? I can build my old installation of 2.071.0 without issue. Those files were converted to glue.d, etc., but obviously the VS build was not updated. Please file a bugzilla issue for it.
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 22:47:34 UTC, Jonathan M Davis wrote: On Tuesday, December 06, 2016 22:13:54 bpr via Digitalmars-d-learn wrote: On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis wrote: Sure, there are folks who would prefer not to have to deal with the GC but throw out the runtime and std lib? You lose out on too much for it to be at all worth it for many folks. At that point, C++11/14/17 looks far more appealing, especially as it continues to improve. It's a counterfactual at this point, but I would guess that if D had left out the GC in 2010 when D2 came out it would have been ahead of C++ in many ways and perhaps would have been able to peel off more C++ programmers and achieve the momentum that Rust appears to have now. Yes, it would be missing some features on account of omitting GC, but D2 -GC in 2010 is still much better than C++ 2011. As C++ absorbs D features, the case for D seems weaker. We get plenty of folks who aren't big C/C++ programmers who are interested in D. Yes, the majority seem to have a C++ background, but we also get folks from C#, python, ruby, etc. It would be nice to see a breakdown. From where I sit, it appears that most of the interest in D is from C++ users, and it doesn't appear that D popularity is rising so much. Any data that belies that sad assessment?
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 22:23:25 UTC, bachmeier wrote: On Tuesday, 6 December 2016 at 22:13:54 UTC, bpr wrote: Those programmers who are comfortable working in a GC-ed language will likely eschew D because D's GC is really not that great. So someone working with Ruby is not going to want to work with D because of GC performance? Ruby programmers are probably not concerned with performance at all ever. It's a slow interpreted language with a GIL. But if you're on a Rails project, that's what you'll use. If I really *want* to use a GC, say I'm writing a server and I believe that a well tuned GC will allow my server to stay alive much longer with less fragmentation, I'll probably skip D and pick Go or maybe (hmmm...) even Java because their GCs have had a lot of engineering effort. I wonder what percentage of Ruby programmers have thought about garbage collection ever. Why would a Ruby or Python programmer unconcerned with performance want to switch to D? I'm sure there are some who would, but I'd imagine they're rare.
DMD 2.072.1 MSVC build seems to be failing
I like to build DMD with MSVC because it complies a little faster than the shipped executable. I downloaded and installed a fresh copy of DMD 2.072.1, opened dmd2/dmd/vcbuild/dmd.sln, and tried to build. I got several compilation errors related to missing files, in particular: 1>c1xx : fatal error C1083: Cannot open source file: '..\glue.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\msc.c': No such file or directory 1>c1xx : fatal error C1083: Cannot open source file: '..\toir.c': No such file or directory I tried fiddling a bit with removing these files from the project, and I could get the first stage of the build to pass, but then I ran into linker errors later on. I don't know enough about how this project is set up, so I figured I'd just ask. Any ideas what's going on? Did the VS project simply not get updated for the new release, or am I missing something in my build procedure? I can build my old installation of 2.071.0 without issue. Thanks!
Re: Large .init for class containing void-initialized struct
On Wednesday, 7 December 2016 at 00:20:11 UTC, Ali Çehreli wrote: tl;dr; go to the TLDR section below. :) [...] struct S { int i = void; double d = void; ubyte[10_000] a = void; } class C { S s = void; // (Same result even without the =void) } void main() { } [...] 04446260 3633 T _D4core4time8Duration13_toStringImplMxFNaNbNfZAya 04505140 3707 T _d_arraysetlengthiT 06681456 00010032 V _D6deneme1C6__initZ Now we have a 10032 byte C.init. Is there a rationale for this or is this an implementation quality issue? Is there a bug already? I could not find one. Also, I failed to find the "= void" documentation e.g. not on the struct spec page. Thank you, Ali Non initialized classes just don't work. Because of the hidden classes fields an initializer is **always** needed. What happens in your example is that the initializer size is sub optimal. A naive make without emplace(): import std.traits, std.c.stdlib; CT make(CT, A...)(A a) { auto memory = malloc(__traits(classInstanceSize, CT)); version(none) emplace!Foo(memory[0..__traits(classInstanceSize, CT)]); static if (__traits(hasMember, CT, "__ctor")) (cast(CT) (memory)).__ctor(a); return cast(CT) memory; } class Foo{void foo(){}} void main() { Foo foo = make!Foo; foo.foo; } crashes with a segfault
[Issue 15485] switch with no case compiles
https://issues.dlang.org/show_bug.cgi?id=15485 Martin Krejcirikchanged: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |INVALID --- Comment #2 from Martin Krejcirik --- This has been an error in dmd since 2.068. GDC bugs should be reported to https://bugzilla.gdcproject.org/ --
Re: [Semi-OT] I don't want to leave this language!
My 2 cents: for most applications, hotspots tend to be in a tiny percentage of the code (ie 90/10 rule where 10% of code accounts for 90% execution time, although my experience in large projects is even more unbalanced) ; throwing away druntime or GC for the whole codebase based on performance concerns amounts to (evil) early optimization. It's far more productive to use a profiler and carefully optimize only the parts that need to be (possibly using betterc, @nogc, ldc and optimized compiler flags for those hotspots only). What could be done better would be to remove friction when linking in shared objects produced by dmd and ldc (eg, extern(D) functions have different number of underscores in dmd vs ldc). On that note, even if dmd is 50x slower than ldc for certain blas mir routines, it's still valuable to have mir supported by dmd as we can always swap in the slow parts during production runs, but we'd benefit with fast compile time during development time, On Tue, Dec 6, 2016 at 3:08 PM, aberba via Digitalmars-d-learn < digitalmars-d-learn@puremagic.com> wrote: > On Tuesday, 6 December 2016 at 22:13:54 UTC, bpr wrote: > >> On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis >> > > I would guess that the vast majority of interest shown in Rust is from >> people who essentially want a better C or C++, with no runtime/GC. So, I >> think Ilya's point is very plausible. D with no GC, but with modules, >> templates, overloading, CTFE, and some other features might have been more >> tempting to the no-GC crowd, which includes many hardcore C++ programmers. >> >> Those programmers who are comfortable working in a GC-ed language will >> likely eschew D because D's GC is really not that great. >> > > > I don't really get the issue with D's GC, Phobos and DRuntime. JavaScript > is really popular and getting really popular everyday (I mean Nodejs). Same > as Python, PHP, Ruby (startups), etc. But they are not exactly betterC. > Most of them don't even give native code speed. > > When using D, I just want to get my app working and running. That is why > more packages (vibe.d, mail, request, mysql-lited, etc) matter to me. The > level you are trying to raise D is way over-kill IMO :). It's good though > for those who need it. But most of us don't judge languages that way. >
Re: Large .init for class containing void-initialized struct
On Wednesday, 7 December 2016 at 00:20:11 UTC, Ali Çehreli wrote: Is there a rationale for this or is this an implementation quality issue? Is there a bug already? I could not find one. Also, I failed to find the "= void" documentation e.g. not on the struct spec page. https://issues.dlang.org/show_bug.cgi?id=11331 https://issues.dlang.org/show_bug.cgi?id=11817 at least. i.e.: known inefficiency, but nobody feels that it is important enough to get to the top of the list.
Re: Large .init for class containing void-initialized struct
On Wednesday, 7 December 2016 at 00:20:11 UTC, Ali Çehreli wrote: tl;dr; go to the TLDR section below. :) I use the following command line to identify large symbols. Given a binary named 'deneme', the following command line (at least on Linux) produces the 30 largest symbols in the 'deneme' binary that actually take space in application's memory: [...] I think this is a bug. Why is the classInit generated at all ? We don't guarantee blit construction of classes./
Large .init for class containing void-initialized struct
tl;dr; go to the TLDR section below. :) I use the following command line to identify large symbols. Given a binary named 'deneme', the following command line (at least on Linux) produces the 30 largest symbols in the 'deneme' binary that actually take space in application's memory: nm --print-size --size-sort --radix=d deneme | tail -30 | grep -v " B " A test program: struct S { int i; double d; ubyte[10_000] a; } void main() { } According to the command line above, the largest symbol in that program is S.init: [...] 04446100 3633 T _D4core4time8Duration13_toStringImplMxFNaNbNfZAya 04504980 3707 T _d_arraysetlengthiT 04511312 00010016 R _D6deneme1S6__initZ So, the S.init object in that binary is 10016 bytes and that makes sense. Now, request S.init not be generated by initializing the members with void: struct S { int i = void; // (Actually, this =void is not required) double d = void; ubyte[10_000] a = void; } void main() { } Great: Now the large S.init is not a part of the binary: (Well, I think it's still in the BSS section but it does not take space in the memory): [...] 04446100 3633 T _D4core4time8Duration13_toStringImplMxFNaNbNfZAya 04504980 3707 T _d_arraysetlengthiT The largest symbol is now something else: _d_arraysetlengthiT. Here comes the trouble... TLDR: Use the void-initialized struct as a class member and that class gets a huge C.init: struct S { int i = void; double d = void; ubyte[10_000] a = void; } class C { S s = void; // (Same result even without the =void) } void main() { } [...] 04446260 3633 T _D4core4time8Duration13_toStringImplMxFNaNbNfZAya 04505140 3707 T _d_arraysetlengthiT 06681456 00010032 V _D6deneme1C6__initZ Now we have a 10032 byte C.init. Is there a rationale for this or is this an implementation quality issue? Is there a bug already? I could not find one. Also, I failed to find the "= void" documentation e.g. not on the struct spec page. Thank you, Ali
[Issue 16952] REG: rdmd --eval='writeln();' keeps failing at almost every release, needs a test
https://issues.dlang.org/show_bug.cgi?id=16952 --- Comment #2 from Timothee Cour--- filed https://issues.dlang.org/show_bug.cgi?id=16953: auto-tester doesn't run rdmd_test (causes a number of regressions in each release) --
[Issue 16953] New: auto-tester doesn't run rdmd_test (causes a number of regressions in each release)
https://issues.dlang.org/show_bug.cgi?id=16953 Issue ID: 16953 Summary: auto-tester doesn't run rdmd_test (causes a number of regressions in each release) Product: D Version: D2 Hardware: x86 OS: All Status: NEW Severity: major Priority: P1 Component: tools Assignee: nob...@puremagic.com Reporter: timothee.co...@gmail.com from https://issues.dlang.org/show_bug.cgi?id=16952: >> Regarding the issue that this stuff comes up again and again: We do have a >> test in rdmd_test.d [1], and it does catch this. I think the problem is that >> the auto-tester doesn't run rdmd_test. and past related bugs: https://issues.dlang.org/show_bug.cgi?id=15175 rdmd --loop and --eval now complain about std.stream deprecation warnings https://issues.dlang.org/show_bug.cgi?id=13288 rdmd --eval fails because it still imports std.metastrings https://issues.dlang.org/show_bug.cgi?id=10352 Regression (2.063): --eval is broken in RDMD --
Re: CTFE Status
On Tuesday, 6 December 2016 at 23:16:16 UTC, Stefan Koch wrote: Yes there is... Thanks.
Re: CTFE Status
On Tuesday, 6 December 2016 at 23:16:16 UTC, Stefan Koch wrote: Yes there is. Create a char[] give it the length of string a + string b; copy string a into the char[] then copy string b into the resulting array. offsetting it by the length of string a. Sounds like you have your concat implementation right there ;)
Re: CTFE Status
On Tuesday, 6 December 2016 at 23:16:16 UTC, Stefan Koch wrote: On Tuesday, 6 December 2016 at 22:44:34 UTC, Nordlöw wrote: On Tuesday, 6 December 2016 at 22:21:59 UTC, Stefan Koch wrote: Go ahead. Many features should already be working correctly. The missing important ones are slices and concat. Ok, great. I was thinking about benchmarking parser generators such as Pegged. Is there another way of concatenating strings than with the builtin ~ operator? Yes there is. Create a char[] give it the length of string a + string b; copy string a into the char[] then copy string b into the resulting array. offsetting it by the length of string a. But keep in mind function calls are unsupported right now. Meaning you have to put everything in one function. However function calls should be working some time soon. I am still in the middle of restructuring how blocks or handled, because right now I will miscompile for-loops with continues in them.
Re: CTFE Status
On Tuesday, 6 December 2016 at 22:44:34 UTC, Nordlöw wrote: On Tuesday, 6 December 2016 at 22:21:59 UTC, Stefan Koch wrote: Go ahead. Many features should already be working correctly. The missing important ones are slices and concat. Ok, great. I was thinking about benchmarking parser generators such as Pegged. Is there another way of concatenating strings than with the builtin ~ operator? Yes there is. Create a char[] give it the length of string a + string b; copy string a into the char[] then copy string b into the resulting array. offsetting it by the length of string a.
[Issue 16952] REG: rdmd --eval='writeln();' keeps failing at almost every release, needs a test
https://issues.dlang.org/show_bug.cgi?id=16952 ag0ae...@gmail.com changed: What|Removed |Added CC||ag0ae...@gmail.com --- Comment #1 from ag0ae...@gmail.com --- (In reply to Timothee Cour from comment #0) > it seems to break at every release, probably a symptom there's a test > missing: > > dmd --version > DMD64 D Compiler v2.072.1 > rdmd '--eval=writeln();' > Error: module cstream is in file 'std/cstream.d' which cannot be read So that people don't go and try to fix this: This specific instance has already been fixed in master. Regarding the issue that this stuff comes up again and again: We do have a test in rdmd_test.d [1], and it does catch this. I think the problem is that the auto-tester doesn't run rdmd_test. [1] https://github.com/dlang/tools/blob/3a32331303e2e62f429b8136ba0de0e9a9294722/rdmd_test.d#L135-L138 --
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 22:13:54 UTC, bpr wrote: On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis I would guess that the vast majority of interest shown in Rust is from people who essentially want a better C or C++, with no runtime/GC. So, I think Ilya's point is very plausible. D with no GC, but with modules, templates, overloading, CTFE, and some other features might have been more tempting to the no-GC crowd, which includes many hardcore C++ programmers. Those programmers who are comfortable working in a GC-ed language will likely eschew D because D's GC is really not that great. I don't really get the issue with D's GC, Phobos and DRuntime. JavaScript is really popular and getting really popular everyday (I mean Nodejs). Same as Python, PHP, Ruby (startups), etc. But they are not exactly betterC. Most of them don't even give native code speed. When using D, I just want to get my app working and running. That is why more packages (vibe.d, mail, request, mysql-lited, etc) matter to me. The level you are trying to raise D is way over-kill IMO :). It's good though for those who need it. But most of us don't judge languages that way.
[Issue 16952] REG: rdmd --eval='writeln();' keeps failing at almost every release, needs a test
https://issues.dlang.org/show_bug.cgi?id=16952 Timothee Courchanged: What|Removed |Added CC||timothee.co...@gmail.com Summary|rdmd --eval='writeln();'|REG: rdmd |keeps failing at every |--eval='writeln();' keeps |release, needs a test |failing at almost every ||release, needs a test Severity|critical|regression --
[Issue 16952] New: rdmd --eval='writeln();' keeps failing at every release, needs a test
https://issues.dlang.org/show_bug.cgi?id=16952 Issue ID: 16952 Summary: rdmd --eval='writeln();' keeps failing at every release, needs a test Product: D Version: D2 Hardware: x86 OS: All Status: NEW Severity: critical Priority: P1 Component: tools Assignee: nob...@puremagic.com Reporter: timothee.co...@gmail.com it seems to break at every release, probably a symptom there's a test missing: dmd --version DMD64 D Compiler v2.072.1 rdmd '--eval=writeln();' Error: module cstream is in file 'std/cstream.d' which cannot be read dmd --version DMD64 D Compiler v2.071.2 rdmd '--eval=writeln();' Error: module syserror is in file 'std/syserror.d' which cannot be read and also earlier versions: https://issues.dlang.org/show_bug.cgi?id=15175 rdmd --loop and --eval now complain about std.stream deprecation warnings https://issues.dlang.org/show_bug.cgi?id=13288 rdmd --eval fails because it still imports std.metastrings https://issues.dlang.org/show_bug.cgi?id=10352 Regression (2.063): --eval is broken in RDMD --
[Issue 8829] std.algorithm.find fails to take advantage of SortedRange
https://issues.dlang.org/show_bug.cgi?id=8829 Andrei Alexandrescuchanged: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --
[Issue 5236] [patch] std.format.formattedRead/unformatValue does not support the raw reading of integer types
https://issues.dlang.org/show_bug.cgi?id=5236 --- Comment #2 from github-bugzi...@puremagic.com --- Commits pushed to master at https://github.com/dlang/phobos https://github.com/dlang/phobos/commit/17d13e20666d388aba04357d98f7ed092b58bccb Issue 5236 - [patch] std.format.formattedRead/unformatValue does not support the raw reading of integer types https://github.com/dlang/phobos/commit/26f4aec45c75e0b6831a1d77614bbca25b5c4068 Issue 5236 - [patch] std.format.formattedRead/unformatValue does not support the raw reading of integer types https://github.com/dlang/phobos/commit/73406665c11c102d4c52e1f12ef9caa1c9ce50ff Issue 5236 - [patch] std.format.formattedRead/unformatValue does not support the raw reading of integer types --
[Issue 8829] std.algorithm.find fails to take advantage of SortedRange
https://issues.dlang.org/show_bug.cgi?id=8829 --- Comment #1 from github-bugzi...@puremagic.com --- Commits pushed to master at https://github.com/dlang/phobos https://github.com/dlang/phobos/commit/8b29f206ae8828cf5656881b16e725850c91b728 Issue 8829 - std.algorithm.find fails to take advantage of SortedRange https://github.com/dlang/phobos/commit/d6519853e1c603169d15c30a1f32aa7c557d6a9d Issue 8829 - std.algorithm.find fails to take advantage of SortedRange https://github.com/dlang/phobos/commit/818d51200754ace7c766c6692a27d95c8e4eafc7 Issue 8829 - std.algorithm.find fails to take advantage of SortedRange https://github.com/dlang/phobos/commit/752b2ca210f8e6f1b7d1c3887f23c1e5a382a85d Issue 8829 - std.algorithm.find fails to take advantage of SortedRange https://github.com/dlang/phobos/commit/e4b82503b8517d533271469797b4a3773ae38527 Issue 8829 - std.algorithm.find fails to take advantage of SortedRange https://github.com/dlang/phobos/commit/98a7c44aafb7e0547fb4b6845e0724597bc81b81 Merge pull request #4907 from RazvanN7/Issue_8828 Issue 8829 - std.algorithm.find fails to take advantage of SortedRange --
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, December 06, 2016 22:13:54 bpr via Digitalmars-d-learn wrote: > On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis > > wrote: > > So, while there are certainly folks who would prefer using D as > > a better C without druntime or Phobos, I think that you're > > seriously overestimating how many folks would be interested in > > that. Certainly, all of the C++ programmers that I've worked > > with professionally would have _zero_ interest in D as a better > > C. > > I would guess that the vast majority of interest shown in Rust is > from people who essentially want a better C or C++, with no > runtime/GC. So, I think Ilya's point is very plausible. D with no > GC, but with modules, templates, overloading, CTFE, and some > other features might have been more tempting to the no-GC crowd, > which includes many hardcore C++ programmers. Sure, there are folks who would prefer not to have to deal with the GC but throw out the runtime and std lib? You lose out on too much for it to be at all worth it for many folks. At that point, C++11/14/17 looks far more appealing, especially as it continues to improve. And @nogc and sane memory use largely solves the GC problem for many programs. There are some places where we need to improve the situation (like with lambdas and the GC), but for most programs, it's totally workable as-is without giving up on all of the features provided by the runtime and Phobos. If you really need absolute pedal-to-the-metal performance and can't afford to ever have the GC stop the world, you still don't need to actually throw away the runtime and std lib. You're just a lot more restricted in what you can do with them. So, tossing out druntime and Phobos entirely seems rather extreme. It may very well make really good sense for a subset of D programs, but I have a hard time believing that it's anything more than a small subset. > Those programmers who are comfortable working in a GC-ed language > will likely eschew D because D's GC is really not that great. We get plenty of folks who aren't big C/C++ programmers who are interested in D. Yes, the majority seem to have a C++ background, but we also get folks from C#, python, ruby, etc. - Jonathan M Davis
Re: CTFE Status
On Tuesday, 6 December 2016 at 22:21:59 UTC, Stefan Koch wrote: Go ahead. Many features should already be working correctly. The missing important ones are slices and concat. Ok, great. I was thinking about benchmarking parser generators such as Pegged. Is there another way of concatenating strings than with the builtin ~ operator?
Tutorial: Form upload in vibe.d
https://aberba.github.io/2016/form-upload-in-vibe-d/
Re: CTFE Status
On Tuesday, 6 December 2016 at 22:21:59 UTC, Stefan Koch wrote: Go ahead. Many features should already be working correctly. This https://github.com/UplinkCoder/dmd/commits/newCTFE I presume?
Re: CTFE Status
On Tuesday, 6 December 2016 at 21:40:47 UTC, Nordlöw wrote: On Tuesday, 6 December 2016 at 16:27:38 UTC, Stefan Koch wrote: I just implemented a bytecode cache, however the bytecode generation is so fast now that we only safe a couple micro seconds by using the cache. It is not noticeable at all :P. I can't wait to try this out. Go ahead. Many features should already be working correctly. The missing important ones are slices and concat.
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 22:13:54 UTC, bpr wrote: Those programmers who are comfortable working in a GC-ed language will likely eschew D because D's GC is really not that great. So someone working with Ruby is not going to want to work with D because of GC performance? I wonder what percentage of Ruby programmers have thought about garbage collection ever. I could see an argument that there are existing frameworks or some other reason to prefer another language, but it is highly unlikely that GC would be the reason that both C++ and Ruby programmers would not want to use D.
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis wrote: So, while there are certainly folks who would prefer using D as a better C without druntime or Phobos, I think that you're seriously overestimating how many folks would be interested in that. Certainly, all of the C++ programmers that I've worked with professionally would have _zero_ interest in D as a better C. I would guess that the vast majority of interest shown in Rust is from people who essentially want a better C or C++, with no runtime/GC. So, I think Ilya's point is very plausible. D with no GC, but with modules, templates, overloading, CTFE, and some other features might have been more tempting to the no-GC crowd, which includes many hardcore C++ programmers. Those programmers who are comfortable working in a GC-ed language will likely eschew D because D's GC is really not that great.
Re: CTFE Status
On Tuesday, 6 December 2016 at 16:27:38 UTC, Stefan Koch wrote: I just implemented a bytecode cache, however the bytecode generation is so fast now that we only safe a couple micro seconds by using the cache. It is not noticeable at all :P. I can't wait to try this out.
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 20:01:38 UTC, Ilya Yaroshenko wrote: On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis wrote: So, while there are certainly folks who would prefer using D as a better C without druntime or Phobos, I think that you're seriously overestimating how many folks would be interested in that. Certainly, all of the C++ programmers that I've worked with professionally would have _zero_ interest in D as a better C. - Jonathan M Davis My experience is completely orthogonal. --Ilya There is a theorem that all programming language discussions on the internet start with the assumption that real programming is exactly the same as the programming done by the author.
Re: Impressed with Appender - Is there design/implementation description?
On Tuesday, 6 December 2016 at 15:29:59 UTC, Jonathan M Davis wrote: On Tuesday, December 06, 2016 13:19:22 Anonymouse via Digitalmars-d-learn wrote: On Tuesday, 6 December 2016 at 10:52:44 UTC, thedeemon wrote: [...] > 2. Up until 4 KB it reallocates when growing, but after 4 KB > the array lives in a larger pool of memory where it can > often grow a lot without reallocating, so in many scenarios > where other allocations do not interfere, the data array of > appender grows in place without copying any data, thanks to > GC.extend() method. I always assumed it kept its own manually allocated array on a malloc heap :O No. The main thing that Appender does is reduce the number of checks required for whether there's room for the array to append in place, because that check is a good chunk of why ~= is expensive for arrays. [...] Thanks everyone for the explanations. I should probably look into my data and see how often I'm reaching the 4kb size triggering GC.extend() use. --Jon
Re: wrong isInputRange design
On Monday, 5 December 2016 at 21:51:28 UTC, pineapple wrote: On Sunday, 4 December 2016 at 11:18:56 UTC, rumbu wrote: --- If you really wanted an `isInputRange` that behaves like you're wanting, it's only a 6 line template that you would have to interject in your code. enum bool isInputRange(T) = is(typeof({ T range = T.init; if(range.empty){} auto element = range.front; range.popFront(); })); https://github.com/pineapplemachine/mach.d My post was about a wrong design decision which generates a compiler error just with one symbol import. If I didn't know how to solve this, I would probably post in the learn section :) After I read the Jonathan's explanations, I understood that it's my fault, one must guess which is the proper way to use a standard library, so I didn't bother to continue the discussion. Anyway, you have a very nice library that just proves my point that a better design is possible: the "asrange" solution for arrays is very elegant. Finally, for my library I decided to completely ignore the range concept, I just check for isIterable and do my processing in a foreach loop.
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 17:00:35 UTC, Jonathan M Davis wrote: So, while there are certainly folks who would prefer using D as a better C without druntime or Phobos, I think that you're seriously overestimating how many folks would be interested in that. Certainly, all of the C++ programmers that I've worked with professionally would have _zero_ interest in D as a better C. - Jonathan M Davis My experience is completely orthogonal. --Ilya
X86-64 calling conventions
I know already that the D calling convention x86-64 is not the same as the one of the corresponding C compiler (register order is reversed on Windows and Linux) - details here: http://forum.dlang.org/post/mailman.1981.1415996395.9932.digitalmar...@puremagic.com But even with extern(C), there is something wrong on x86-64 when passing arrays as parameters. In x86-32, the D calling convention is clear for arrays (even that's not documented, credits go to the std.bigint creator - Don Clungston): void foo(uint[] x, uint p1, uint p2) will be translated to push x.ptr; push x.length; push p1; push p2; call foo; In a "naked" asm context, it's easy to access parameters p2 = [ESP + 4] p1 = [ESP + 8] x.length = [ESP + 12] x.ptr = [ESP + 16] --- Now, on x86-64 (Windows), following the same schema, I expected something like this: mov RCX, x.ptr; mov RDX, x.length; mov R8, p1; mov R9, p2; (standard Windows x86-64 calling convention) Instead, the compiler generates something very strange: mov R8, p2 //somehow ok, third parameter (even I expected to be the fourth) mov RDX, p1 //somehow ok, second parameter (even I expected to be the third) mov [RBP-0x10], x.length //why? mov [RBP-0x08], x.ptr //why? lea RCX, [RBP-0x10] //pointer to x.length? Finally we have: p2 = R8 p1 = RDX x.length = [RCX] x.ptr = ? Now the questions: 1. Is there any intention to correct the registry reversing behaviour for extern(D)? 2. Will someone fully document the ABI for x86-32 and x86-64? 3. How do I find in a naked asm context on x86-64 the starting address of an array passed as parameter? Thanks.
Re: What's the oldest Mac targeted by D programmers?
On Tuesday, 6 December 2016 at 14:28:37 UTC, Walter Bright wrote: dmd generates SSE code for OSX 32. Unfortunately Windows + 32-bit is disproportionately important for my line of work.
Re: What's the oldest Mac targeted by D programmers?
On Tuesday, 6 December 2016 at 16:00:43 UTC, Nick Sabalausky wrote: "Latest is greatest" has become such a sacred cow in the tech sector that far too few developers these days understand basic common-sense logic like that. I guess it's all about if you control the target hardware. It would be different for scientific software.
Re: CTFE Status
On 12/06/2016 11:27 AM, Stefan Koch wrote: On Monday, 5 December 2016 at 18:47:13 UTC, Stefan Koch wrote: On Monday, 5 December 2016 at 16:47:33 UTC, Andrei Alexandrescu wrote: On 12/05/2016 11:28 AM, Stefan Koch wrote: It looks like the performance wins brought by the new ctfe engine might be higher then I predicted. That's awesome!! -- Andrei After discovering this performance bottleneck I have now changed my mind about how I will tackle concatenation of arrays and strings in particular. All con-cat operations should be done as intrinsic calls. Because tight cooperation with the CTFE-Memory-Management-Subsystem is needed. I just implemented a bytecode cache, however the bytecode generation is so fast now that we only safe a couple micro seconds by using the cache. It is not noticeable at all :P. Just give us time :o). -- Andrei
[Issue 16951] New: trying to call opCall with alias opCall this, without parentesis fails to compile, when used as a single statement
https://issues.dlang.org/show_bug.cgi?id=16951 Issue ID: 16951 Summary: trying to call opCall with alias opCall this, without parentesis fails to compile, when used as a single statement Product: D Version: D2 Hardware: x86_64 OS: Linux Status: NEW Severity: enhancement Priority: P1 Component: dmd Assignee: nob...@puremagic.com Reporter: var.spool.mail...@gmail.com struct A { alias opCall this; int opCall() { return 0; } } void main() { A a; // a; // Error: var has no effect in expression (a) // some working examples but with some issues auto a1 = a; // calls copy constructor int a2 = a; // calls opCall a.writeln;// toString a.someIntFun; // opCall } those issues can be partialy solved by adding @property to opCall struct A { alias opCall this; @property int opCall() { return 0; } } void main() { A a; // a; // still Error: var has no effect in expression (a) auto a1 = a; // copy constructor int a2 = a; // opCall a.writeln;// opCall a.someIntFun; // opCall } dont know if its a bug or not but would be nice if the commented line would compile. --
Re: How about a bounty for a new windows installer using inno setup ?
On 12/6/2016 12:21 AM, Thomas Mader wrote: You can also create a WiX installer out of an InnoSetup installer. I think it's more important to decide upon the feature set, readability and the time needed to build an installer. Have you experience with both? I only have experience with NSIS and InnoSetup and in InnoSetup the feature set for Windows is really good and the readability is good. I started out by using InstallShield some years ago and got battle scars there. I don't recommend that. I used NSIS a little because a company we partnered with required it but I'm no authority on NSIS or Inno. It really depends on how complicated your particular install is and where you expect it to go in the future. If you're just copying a few files then anything will work. I don't mean to make too big a deal out of it if the requirements are really simple. Personally I think it's better in the long run to generate an MSI for several reasons you can probably look up yourself---security, ability to rollback (installation is a transaction), appears in Programs and Features, transforms, etc. I wouldn't advise doing the coding part externally in D this makes things much more complicated than it should be. Stick with what's supported by the tool. I'm not suggesting you necessarily use D together with something like NSIS. But you do want to have a one-button automated build process, not just for convenience but for repeatability. That's important. Soon you will want to get away from the tool's own GUI and run things programmatically. After having not-so-good experiences with InstallShield I looked at things like SCons and msbuild, which was just coming out at the time. (This was a while ago). I tried msbuild but it didn't have modules to support many of the things I needed to do. Things like code-signing with a verisign signature, injecting data and files into exe resources, etc. Fortunately you can build your own custom modules using C# which is what I tried. But the process of transferring variables back and forth from the script to C# for every custom module was painful and I thought, "if I just do this all in C# it will be much easier". So I switched. I used C# not only to call on the WiX tools to run them, but to easily manipulate pathname and filename strings, which were different because I needed to build different configurations for different customer companies. And I also needed to build different combinations of language localization. I could use .NET to build a nice GUI for selecting configurations, and C# to call the Windows API when needed, move files around, anything. You may also run into issues when you need to do complicated updates and there are already earlier versions in the field. You may want to remove features, but your installer has to both update existing users in addition to supporting new users. It's hard to predict the future though so I won't say much about it except that it helps to have a more powerful tool when you run into such situations. The scripting-type tools are tempting because they're easy and no one wants to spend any time on installers. It's usually something that people hope to just slap on at the end and it often gets underestimated. But as I said, maybe it IS easy if you're just copying files and you will only ever have one configuration. So it depends. Jim
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, December 06, 2016 13:36:20 Ilya Yaroshenko via Digitalmars-d- learn wrote: > On Tuesday, 6 December 2016 at 13:02:16 UTC, Andrei Alexandrescu > > wrote: > > On 12/6/16 3:28 AM, Ilya Yaroshenko wrote: > >> On Tuesday, 6 December 2016 at 08:14:17 UTC, Andrea Fontana > >> > >> wrote: > >>> On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko > >>> Phobos/Druntime are pretty good for a lot of projects. > >> > >> In theory > > > > And what seem to be the issues in practice with code that is > > not highly specialized? -- Andrei > > If code is not highly specialized there is no reason to spent > resources to use C/C++/D. A company will be happy with Python, > Java, C#, Go and Swift. If one need to have C/C++ programming > level he can not use D because DRuntime. Only a subset of D can > be used. And current problem that we have not BetterC paradigm in > D specification. So, only crazy companies will consider D for > large projects. Current D is successful in small console text > routines. > > If a system PL can not be used as C for highly specialized code, > it is not a real system PL. > > DRuntime and Phobos is going to compete with Java and Go. It is > suicide for D, IMHO. In other hand, BetterC is a direction where > D can be populated among professionals and replace C/C++. While I am quite sure that there are use cases where folks would be unhappy with druntime and Phobos and want to avoid them, there are definitely companies using D with druntime and Phobos right now, and personally, I've never worked at a company where anything about druntime or Phobos would have been a showstopper in switching to D. The showstopper would be in convincing them to use a new language rather than one that they were already familiar with. C++ works for them, and they're not interested in switching. And honestly, pushing for a subset of D that didn't have the functionality in druntime and Phobos would just make it an even harder sell. At that point, they're _really_ not interested in using D. C++11/14/17 quickly and easily wins out in that fight. So, while there are certainly folks who would prefer using D as a better C without druntime or Phobos, I think that you're seriously overestimating how many folks would be interested in that. Certainly, all of the C++ programmers that I've worked with professionally would have _zero_ interest in D as a better C. - Jonathan M Davis
Re: CTFE Status
On Monday, 5 December 2016 at 18:47:13 UTC, Stefan Koch wrote: On Monday, 5 December 2016 at 16:47:33 UTC, Andrei Alexandrescu wrote: On 12/05/2016 11:28 AM, Stefan Koch wrote: It looks like the performance wins brought by the new ctfe engine might be higher then I predicted. That's awesome!! -- Andrei After discovering this performance bottleneck I have now changed my mind about how I will tackle concatenation of arrays and strings in particular. All con-cat operations should be done as intrinsic calls. Because tight cooperation with the CTFE-Memory-Management-Subsystem is needed. I just implemented a bytecode cache, however the bytecode generation is so fast now that we only safe a couple micro seconds by using the cache. It is not noticeable at all :P.
Re: What's the oldest Mac targeted by D programmers?
On 12/06/2016 08:11 AM, Guillaume Piolat wrote: optimizing for the users which already have the fastest machines is not going to make a difference while optimizing for the lower end does. Very wise. "Latest is greatest" has become such a sacred cow in the tech sector that far too few developers these days understand basic common-sense logic like that.
Re: wrong isInputRange design
On Monday, 5 December 2016 at 21:51:28 UTC, pineapple wrote: https://github.com/pineapplemachine/mach.d I love that a few of the subfolders have a Readme.md so that I don't have to dig in to anything to get an overview.
Re: Impressed with Appender - Is there design/implementation description?
On Tuesday, December 06, 2016 13:19:22 Anonymouse via Digitalmars-d-learn wrote: > On Tuesday, 6 December 2016 at 10:52:44 UTC, thedeemon wrote: > > It's rather simple, just take a look at its source code in > > std.array. > > It's an ordinary linear array partially filled with your data. > > [...] > > > 2. Up until 4 KB it reallocates when growing, but after 4 KB > > the array lives in a larger pool of memory where it can often > > grow a lot without reallocating, so in many scenarios where > > other allocations do not interfere, the data array of appender > > grows in place without copying any data, thanks to GC.extend() > > method. > > I always assumed it kept its own manually allocated array on a > malloc heap :O No. The main thing that Appender does is reduce the number of checks required for whether there's room for the array to append in place, because that check is a good chunk of why ~= is expensive for arrays. And using ~= isn't all that bad. It's just that it adds up, and when you're clearly going to be doing a lot of appending when you first create a dynamic array, it's just more efficient to use Appender and avoid the extra overhead that ~= has. What Appender does with put is actually pretty close to what druntime does with ~=. It's just that Appender can just look at the member variable to know the capacity of the dynamic array, whereas ~= has to look it up by looking at the data for the memory block that the dynamic array currently points to. All Appender really is is a way to make appending to dynamic arrays more efficient. It doesn't do anything magical with data structures. At present, this is what it's member variables look like private alias T = ElementEncodingType!A; private struct Data { size_t capacity; Unqual!T[] arr; bool canExtend = false; } private Data* _data; The magic is in that extra capacity variable, but otherwise, you're basically dealing with the same array appending you get out of ~=. > Hence a practice of using .dup and .idup on the .data member when > you're happy with the result? That isn't necessary. I don't think that I ever do it. If you want to keep appending with Appender, and you want a dynamic array that has its own memory to grow into without being affected by Appender, then it would make sense to use dup/idup, but that's true any time that you append to a dynamic array. And you might want dup/idup to get a dynamic array with a different type of mutability than the dynamic array in the Appender just like you might want with a naked dynamic array, but since you define the mutability with Appender, you'd usually just change the array type that you instantiated Appender with. - Jonathan M Davis
Re: What's the oldest Mac targeted by D programmers?
On 12/6/2016 5:11 AM, Guillaume Piolat wrote: No opinion. My whole segment runs on SSE. Zero interest in AVX here, because optimizing for the users which already have the fastest machines is not going to make a difference while optimizing for the lower end does. Hence it's critically important that 32-bit builds use SSE instead of FPU (because denormals) but that's another question and already addressed by LDC. For DMD I'd more interested in yet improving raw compile-time speed, quality, and working 32-bit SSE intrinsics in core.intrinsic (did that changed? D_SIMD wasn't here in 32-bit). dmd generates SSE code for OSX 32.
[Issue 16758] Variant.opIndex result not modified after opAssign
https://issues.dlang.org/show_bug.cgi?id=16758 github-bugzi...@puremagic.com changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --
[Issue 16758] Variant.opIndex result not modified after opAssign
https://issues.dlang.org/show_bug.cgi?id=16758 --- Comment #1 from github-bugzi...@puremagic.com --- Commits pushed to master at https://github.com/dlang/phobos https://github.com/dlang/phobos/commit/cdd2acb0d58b38659260348d4fb78d2aba6db51c Fix Issue 16758 - Variant.opIndex result not modified after opAssign Add VariantN.opIndexOpAssign. Also undocument & tweak test for const(Variant).opIndex. https://github.com/dlang/phobos/commit/94d3786d6995dbc2fafdba1408dd6d7974d91625 Merge pull request #4923 from ntrel/variant-opIndexOpAssign Fix Issue 16758 - Variant.opIndex result not modified after opAssign --
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 13:02:16 UTC, Andrei Alexandrescu wrote: On 12/6/16 3:28 AM, Ilya Yaroshenko wrote: On Tuesday, 6 December 2016 at 08:14:17 UTC, Andrea Fontana wrote: On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko Phobos/Druntime are pretty good for a lot of projects. In theory And what seem to be the issues in practice with code that is not highly specialized? -- Andrei If code is not highly specialized there is no reason to spent resources to use C/C++/D. A company will be happy with Python, Java, C#, Go and Swift. If one need to have C/C++ programming level he can not use D because DRuntime. Only a subset of D can be used. And current problem that we have not BetterC paradigm in D specification. So, only crazy companies will consider D for large projects. Current D is successful in small console text routines. If a system PL can not be used as C for highly specialized code, it is not a real system PL. DRuntime and Phobos is going to compete with Java and Go. It is suicide for D, IMHO. In other hand, BetterC is a direction where D can be populated among professionals and replace C/C++. Ilya
Re: Impressed with Appender - Is there design/implementation description?
On Tuesday, 6 December 2016 at 10:52:44 UTC, thedeemon wrote: It's rather simple, just take a look at its source code in std.array. It's an ordinary linear array partially filled with your data. [...] 2. Up until 4 KB it reallocates when growing, but after 4 KB the array lives in a larger pool of memory where it can often grow a lot without reallocating, so in many scenarios where other allocations do not interfere, the data array of appender grows in place without copying any data, thanks to GC.extend() method. I always assumed it kept its own manually allocated array on a malloc heap :O Hence a practice of using .dup and .idup on the .data member when you're happy with the result?
Re: What's the oldest Mac targeted by D programmers?
On Tuesday, 6 December 2016 at 09:12:16 UTC, Walter Bright wrote: This leads to the question is AVX support a minimum configuration for Macs? Can AVX instruction generation become the default for OSX? No opinion. My whole segment runs on SSE. Zero interest in AVX here, because optimizing for the users which already have the fastest machines is not going to make a difference while optimizing for the lower end does. Hence it's critically important that 32-bit builds use SSE instead of FPU (because denormals) but that's another question and already addressed by LDC. For DMD I'd more interested in yet improving raw compile-time speed, quality, and working 32-bit SSE intrinsics in core.intrinsic (did that changed? D_SIMD wasn't here in 32-bit).
Re: [Semi-OT] I don't want to leave this language!
On 12/5/16 3:49 PM, e-y-e wrote: If you don't mind me saying, I think Mir could be one of the best things for the future of D (along with LDC) and I'd be glad to help it on its way. Yes, Mir is awesome! I keep on thinking of ways to make it better supported by the language and infra. -- Andrei
Re: [Semi-OT] I don't want to leave this language!
On 12/6/16 3:28 AM, Ilya Yaroshenko wrote: On Tuesday, 6 December 2016 at 08:14:17 UTC, Andrea Fontana wrote: On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko Phobos/Druntime are pretty good for a lot of projects. In theory And what seem to be the issues in practice with code that is not highly specialized? -- Andrei
Re: Impressed with Appender - Is there design/implementation description?
On Sunday, 4 December 2016 at 20:03:37 UTC, Jon Degenhardt wrote: I've been using Appender quite a bit recently, typically when I need append-only arrays with variable and unknown final sizes. I had been prepared to write a custom data structure when I started using it with large amounts of data, but very nicely this has not surfaced as a need. Appender has held up quite well. I haven't actually benchmarked it against competing data structures, nor have I studied the implementation. I'd be very interested in understanding the design and how it compares to other data structures. Are there any write-ups or articles describing it? --Jon It's rather simple, just take a look at its source code in std.array. It's an ordinary linear array partially filled with your data. When your data fills it up, it gets resized to make more space. Just two interesting points: 1. When it needs to grow it uses a formula like k = 1 + 10 / log2(newsize) for the growth factor (but with a limit of k <= 2), which means up to 16 KB it doubles the size each time, then tries to grow by a factor of 2/3, then by lower and lower factors. 2. Up until 4 KB it reallocates when growing, but after 4 KB the array lives in a larger pool of memory where it can often grow a lot without reallocating, so in many scenarios where other allocations do not interfere, the data array of appender grows in place without copying any data, thanks to GC.extend() method.
Re: Set Intersection of SortedRanges
On Monday, 5 December 2016 at 22:10:34 UTC, Nordlöw wrote: Has anybody already done this? This gives good guidance on three different approaches. http://stackoverflow.com/questions/2400157/the-intersection-of-two-sorted-arrays/4601106#4601106 Luckily we already have galloping search in Phobos :) I'll do a couple of benchmarks.
Re: What's the oldest Mac targeted by D programmers?
On 12/6/2016 1:31 AM, Johan Engelen wrote: AVX2 is not listed in machdep.cpu.features, at least not on my machine; you need to look in the machdep.cpu.leaf7_features list. https://csharpmulticore.blogspot.nl/2014/12/how-to-check-intel-avx2-support-on-mac-os-x-haswell.html Bummer. Seems like I have nothing that does AVX2. Worse, my motherboard will not support the AMD Carrizo processor (the only AMD AVX2 one) so upgrading means a new computer. Grumble, grumble.
Re: What's the oldest Mac targeted by D programmers?
On Tuesday, 6 December 2016 at 09:12:16 UTC, Walter Bright wrote: I've been experimenting with code generation in DMD for the AVX instruction set, in particular the replacement of SSE 16 byte vectors with AVX 16 byte vectors (no, not the 32 byte ones!). In my experiments on my machines, two of them support the AVX instruction set, one of which is my Mac Mini, as determined by this program: import core.stdc.stdio; import core.cpuid; void main() { printf("%d %d\n", core.cpuid.avx, core.cpuid.avx2); } or by running: sysctl -a | grep machdep.cpu.features and looking for AVX1.0. My Mac Mini sez: ~> sysctl -a | grep machdep.cpu.features machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ DTES64 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 SSE4.2 xAPIC POPCNT AES PCID XSAVE OSXSAVE TSCTMR AVX1.0 My older MM clearly does not support AVX2, though. This leads to the question is AVX support a minimum configuration for Macs? Can AVX instruction generation become the default for OSX? No. Avx arrived in sandy bridge, early 2011. The first sandy bridge processors to land in Macs were shortly after that. macOS Sierra supports hardware back to late 2009 and the OS X versions we support work on even earlier hardware. See everymac.com for apple hardware details. As far I can tell, -march= / -mcpu= options with native as a valid specification is the standard, sane way to deal with this.
Re: What's the oldest Mac targeted by D programmers?
On Tuesday, 6 December 2016 at 09:12:16 UTC, Walter Bright wrote: My Mac Mini sez: ~> sysctl -a | grep machdep.cpu.features machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ DTES64 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 SSE4.2 xAPIC POPCNT AES PCID XSAVE OSXSAVE TSCTMR AVX1.0 My older MM clearly does not support AVX2, though. AVX2 is not listed in machdep.cpu.features, at least not on my machine; you need to look in the machdep.cpu.leaf7_features list. https://csharpmulticore.blogspot.nl/2014/12/how-to-check-intel-avx2-support-on-mac-os-x-haswell.html -Johan
What's the oldest Mac targeted by D programmers?
I've been experimenting with code generation in DMD for the AVX instruction set, in particular the replacement of SSE 16 byte vectors with AVX 16 byte vectors (no, not the 32 byte ones!). In my experiments on my machines, two of them support the AVX instruction set, one of which is my Mac Mini, as determined by this program: import core.stdc.stdio; import core.cpuid; void main() { printf("%d %d\n", core.cpuid.avx, core.cpuid.avx2); } or by running: sysctl -a | grep machdep.cpu.features and looking for AVX1.0. My Mac Mini sez: ~> sysctl -a | grep machdep.cpu.features machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ DTES64 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 SSE4.2 xAPIC POPCNT AES PCID XSAVE OSXSAVE TSCTMR AVX1.0 My older MM clearly does not support AVX2, though. This leads to the question is AVX support a minimum configuration for Macs? Can AVX instruction generation become the default for OSX?
Re: D with CygWin
So it was not finish :-( mkdir("C:\\cygwin\\home\\undefer\\TEST") is working mkdir("/home/undefer/TEST") creates directory and hangs, doesn't pass control to the next instruction. DirEntry("/") works in my simple program, but doesn't work in big program.. windbg, gdb doen't help at all. It is failure :-(
Re: [Semi-OT] I don't want to leave this language!
On Monday, 5 December 2016 at 20:49:50 UTC, e-y-e wrote: On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko wrote: [...] You know from the 15th December I will have a month of free time, and I would love to get myself up to speed with Mir to contribute to it. If you don't mind me saying, I think Mir could be one of the best things for the future of D (along with LDC) and I'd be glad to help it on its way. Awesome! The main directions are: 1. stdC++ analogs implementation in betterC Dlang subset 2. betterC analogs of existing Phobos modules. We can reuse DRuntime / Phobos code for initial commits. 3. Various numeric / sci software. 4. GPU algorithms. This require dcompute to be a part of LDC. Thank you, Ilya
Re: [Semi-OT] I don't want to leave this language!
On Tuesday, 6 December 2016 at 08:14:17 UTC, Andrea Fontana wrote: On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko Phobos/Druntime are pretty good for a lot of projects. In theory
Re: How about a bounty for a new windows installer using inno setup ?
On Monday, 5 December 2016 at 19:33:33 UTC, Jim Hewes wrote: On 12/5/2016 3:19 AM, Kjartan F. Kvamme wrote: On Monday, 5 December 2016 at 09:24:59 UTC, Basile B. wrote: How about a bounty for a new windows installer using inno setup ? There are several issues related to the nsis-based windows installer (even on bugzilla). The problem that happened last Fall with a virus false detection may happen again. "Braddr" proposed to handle digital signatures in case it would involve payment. Programming an installer is a small job but it has a long term impact on the user experience. Worth 100€ imo. Any particular reason to use Inno Setup over for example Wix Toolset? In my last job I worked on installers (which I didn't like but someone had to do it.) I recommend WiX over Inno. The main reason is that WiX produces an MSI and Inno doesn't. An MSI is just a data file, not an executable, and is thus better for security. I normally wrapped the MSI in a bootstrap exe. But we had one customer that was part of the government and wouldn't accept anything but an MSI. If you want, you can generate the XML with a program. I just didn't because I figured it was easier to modify if you can directly see the XML. My install builder was actually a combination of C# and WiX. I never found scripts to be flexible enough and it's just one more language to know. Jim You can also create a WiX installer out of an InnoSetup installer. I think it's more important to decide upon the feature set, readability and the time needed to build an installer. Have you experience with both? I only have experience with NSIS and InnoSetup and in InnoSetup the feature set for Windows is really good and the readability is good. I wouldn't advise doing the coding part externally in D this makes things much more complicated than it should be. Stick with what's supported by the tool.
Re: [Semi-OT] I don't want to leave this language!
On Monday, 5 December 2016 at 20:25:00 UTC, Ilya Yaroshenko wrote: Good D code should be nothrow, @nogc, and betterC. BetterC means that it must not require DRuntime to link and to start. I started Mir as scientific/numeric project, but it is going to be a replacement for Phobos to use D instead/with of C/C++. Completly disagree. You're speaking about scientific projects, maybe. Phobos/Druntime are pretty good for a lot of projects. Andrea