Re: mysql-native: API Refresh RC
Am 30.01.2017 um 03:56 schrieb Nick Sabalausky: (...) So take a look, let me know if there's any big issues with it. If all looks good, this will soon be released as mysql-native v0.2.0. What about directly going for 1.0.0? At least after it has gotten enough real-world exposure, I'd say that the first API overhaul is a good opportunity for that.
Re: unDE 0.2.0 - unusual command line and keybar
On Sunday, 29 January 2017 at 23:57:30 UTC, FreeSlave wrote: On Sunday, 29 January 2017 at 19:00:30 UTC, unDEFER wrote: Very interesting concept (Probably it's not new, but I never actually used file managers like this). It looks you put much love and effort in it. Damn, you even made packages for many different Linux distros! (I'm always lazy when things come to packaging, so I'm just amazed that you made so many). Thank you Why did you not give a link to github repo? I had to search it myself https://github.com/unDEFER/unde Because it is only mirror and I update it only on releases. How can I build and run unde myself? I built it via dub build and installed ttf-ancient-fonts, but it still throws exception when I try to run ./unde: object.Exception@source/unde/font.d(714): TTF_OpenFont: Couldn't open /usr/share/fonts/truetype/ancient-scripts/Symbola_hint.ttf Look source/unde/font.d, there is many "versions" where may be installed these fonts. If you find your paths you can use 'dub build -c "Version"'. Else you can add own or make symbolic links. Also you can say your distro and maybe the next time I will make package for you.
mysql-native: API Refresh RC
I've been working on a big refresh of mysql-native's API, to take care of various issues that have appeared with it. It involves some major breaking changes (although I've tried to keep old interfaces around for the moment, but marked deprecated), so I wanted to post it before committing to it so those interested have a change to take a look, give feedback, catch problems, etc. Summary of these changes: API overhauled for better safety, reliability and ease-of-use. Deprecated and replaced entire Command struct with better design. Better handling of null. Various bugs fixed and more rigorously tested. -- For right now, the changes are in a separate fork, here: https://github.com/Abscissa/mysql-native-experimental The readme there has sample code and an overview of the new interface. Changelog: https://github.com/Abscissa/mysql-native-experimental/blob/master/CHANGELOG.md API ref: http://semitwist.com/mysql-native-docs/v0.2.0-preview1 --- So take a look, let me know if there's any big issues with it. If all looks good, this will soon be released as mysql-native v0.2.0.
Re: Release D 2.073.0
On Monday, 30 January 2017 at 01:15:52 UTC, Dicebot wrote: On 01/30/2017 12:38 AM, Walter Bright wrote: ... Please, don't waste your time. You mentioned being curious about what is wrong with that PR - I have explained. Let's just stop here before you write another 20 posts presuming that I only disagree with your development methodology because I don't understand it. Walter created an entire language and a community around it. Can you, please, share with us how your accomplishments give any importance to whatever your disagreement is with him? All that is visible, here is you protest everything, take any opportunity to verbally abuse everyone and make no contribution. Thanks.
Re: Release D 2.073.0
On 01/30/2017 12:38 AM, Walter Bright wrote: > ... Please, don't waste your time. You mentioned being curious about what is wrong with that PR - I have explained. Let's just stop here before you write another 20 posts presuming that I only disagree with your development methodology because I don't understand it. signature.asc Description: OpenPGP digital signature
Re: Release D 2.073.0
On 1/29/2017 4:18 AM, Dicebot wrote: On Friday, 27 January 2017 at 19:12:37 UTC, Walter Bright wrote: On 1/27/2017 3:12 AM, Dicebot wrote: And also stuff like https://github.com/dlang/druntime/pull/1740 I'm curious what is wrong with that? You have been pushing for premature merged of `return scope` under a premise that it will be hidden behind a switch and won't affect anyone yet. Now you rush to adjust druntime to use it Drilling down in Phobos to make it work with -dip1000, at the bottom are usually calls to core.stdc.*. Without annotating those functions, all progress on -dip1000 comes to a halt. -dip1000 is designed so that non-dip1000 code can call dip1000 code. Adding these annotations should be benign to existing code. Nothing failed after adding these annotations, either in Phobos which is not (yet) dip1000, nor any of the projects tested, nor any of the test suite. and require the same from any druntime contributors. Since druntime is now compiled with -dip1000, and future druntime contributions are not existing code, that is a reasonable expectation. On the same premise, I am working to get Phobos compatible with -dip1000. Without that compatibility, -dip1000 is pretty much unusable. It's analogous to druntime and phobos being officially compiled with -dip25, which has been the case for some time now.
Re: unDE 0.2.0 - unusual command line and keybar
On Sunday, 29 January 2017 at 19:00:30 UTC, unDEFER wrote: unDE's not DE which in the future must be replacement for all programs in OS. But today is very original file manager, image and text viewer and (what discovered with 0.2.0 version) command line and keybar. More information: http://unde.sourceforge.net/en/ch25.html Video with English subtitles: https://youtu.be/XY7Jegxq6vg Very interesting concept (Probably it's not new, but I never actually used file managers like this). It looks you put much love and effort in it. Damn, you even made packages for many different Linux distros! (I'm always lazy when things come to packaging, so I'm just amazed that you made so many). Why did you not give a link to github repo? I had to search it myself https://github.com/unDEFER/unde How can I build and run unde myself? I built it via dub build and installed ttf-ancient-fonts, but it still throws exception when I try to run ./unde: object.Exception@source/unde/font.d(714): TTF_OpenFont: Couldn't open /usr/share/fonts/truetype/ancient-scripts/Symbola_hint.ttf
unDE 0.2.0 - unusual command line and keybar
unDE's not DE which in the future must be replacement for all programs in OS. But today is very original file manager, image and text viewer and (what discovered with 0.2.0 version) command line and keybar. More information: http://unde.sourceforge.net/en/ch25.html Video with English subtitles: https://youtu.be/XY7Jegxq6vg
Re: columnar struct storage template
On Sunday, 29 January 2017 at 14:16:37 UTC, ezneh wrote: Here : foreach(uint u; FooCol.j) writef("%s ", u); writef(\n); I see, but apparently this isn't the only issue. I replaced the whole unittest with this, and it still doesn't compile: import std.stdio; void main() { struct MyStruct { uint i; float j; } MyStruct data; Columns!(MyStruct) MyCols; } Problems this time were in the AfterLast function, which I modified like this: string AfterLast(string s, char c) { foreach_reverse(int i, char d; s) if(c==d) return s[i+1..$]; return s; } However there are still some errors which I don´t know how to fix: columns1.d(45): Error: cannot implicitly convert expression (this) of type Columns!(MyStruct) to Columns!(MyStruct)* columns1.d(63): Error: template instance columns1.Columns!(MyStruct) error instantiating Line 63 is the declaration of MyCols, and line 45 would be thisÑ At opIndex(uint i) { return At(this,i); }
Re: columnar struct storage template
On Sunday, 29 January 2017 at 14:00:37 UTC, Nestor wrote: On Sunday, 29 January 2017 at 13:45:12 UTC, Adam D. Ruppe wrote: On Sunday, 29 January 2017 at 13:09:18 UTC, Nestor wrote: This no longer compiles on recent versions of DMD :( Trivial fix though, it just needs to say "\n" instead of plain \n (it is missing quotes, D used to allow that for whatever stupid reason and has since been fixed). Actually, tbh, I'm surprised it isn't a bigger fix needed given that that post is from 9 years ago! Where is it missing the quotes? I tried fixing it like this, but it still doesn't compile: "(){return members["~j.stringof~"];}\"\n\"" Here : foreach(uint u; FooCol.j) writef("%s ", u); writef(\n);
Re: columnar struct storage template
On Sunday, 29 January 2017 at 13:45:12 UTC, Adam D. Ruppe wrote: On Sunday, 29 January 2017 at 13:09:18 UTC, Nestor wrote: This no longer compiles on recent versions of DMD :( Trivial fix though, it just needs to say "\n" instead of plain \n (it is missing quotes, D used to allow that for whatever stupid reason and has since been fixed). Actually, tbh, I'm surprised it isn't a bigger fix needed given that that post is from 9 years ago! Where is it missing the quotes? I tried fixing it like this, but it still doesn't compile: "(){return members["~j.stringof~"];}\"\n\""
Re: columnar struct storage template
On Sunday, 29 January 2017 at 13:09:18 UTC, Nestor wrote: This no longer compiles on recent versions of DMD :( Trivial fix though, it just needs to say "\n" instead of plain \n (it is missing quotes, D used to allow that for whatever stupid reason and has since been fixed). Actually, tbh, I'm surprised it isn't a bigger fix needed given that that post is from 9 years ago!
Re: columnar struct storage template
On Saturday, 13 December 2008 at 21:13:52 UTC, BCS wrote: If you have an array of structs that is primarily accessed by columns (the same member in each item) rather than by rows (all members in a given item) it can be faster to store the members each in there own array. The Columns template automates this for POD struct types. http://www.dsource.org/projects/scrapple/browser/trunk/columns/columns.d This no longer compiles on recent versions of DMD :(
Re: Release D 2.073.0
On Friday, 27 January 2017 at 19:12:37 UTC, Walter Bright wrote: On 1/27/2017 3:12 AM, Dicebot wrote: And also stuff like https://github.com/dlang/druntime/pull/1740 I'm curious what is wrong with that? You have been pushing for premature merged of `return scope` under a premise that it will be hidden behind a switch and won't affect anyone yet. Now you rush to adjust druntime to use it and require the same from any druntime contributors.
Re: Release D 2.073.0
Continuing on a new thread because this is getting kinda off-topic. http://forum.dlang.org/post/jhtvuvhxsayjatsdb...@forum.dlang.org