Difference between revisions of "Wish list"

From D Wiki
Jump to: navigation, search
m (std.encoding)
(add an overview of existing community libs)
Line 1: Line 1:
'''Warning''': This wish list is partially outdated and needs some updates. For the time being head over to [[Review Queue]] to see the status of current additions.
 
 
 
== Wish list ==
 
== Wish list ==
  
Line 30: Line 28:
 
=== std.database ===
 
=== std.database ===
 
This is a large, complicated project because we're trying to define a common API for a variety of RDBMSs. Again, it's more a design problem than an implementation problem. There is at least one current attempt at a unified interface: [https://github.com/buggins/ddbc DDBC]
 
This is a large, complicated project because we're trying to define a common API for a variety of RDBMSs. Again, it's more a design problem than an implementation problem. There is at least one current attempt at a unified interface: [https://github.com/buggins/ddbc DDBC]
 +
(See also: [Database_Libraries]).
  
 
=== Parallel algorithms ===
 
=== Parallel algorithms ===
Line 44: Line 43:
  
 
[http://dlang.org/phobos/etc_c_zlib.html etc.c.zlib] is a wrapper around the low-level bindings to the C library Zlib which has to be included in the D release. Apart from the licensing issues, it also means that maintaining and extending this library is quite hard.
 
[http://dlang.org/phobos/etc_c_zlib.html etc.c.zlib] is a wrapper around the low-level bindings to the C library Zlib which has to be included in the D release. Apart from the licensing issues, it also means that maintaining and extending this library is quite hard.
 +
The best existing library is [https://github.com/rcythr/archive archive].
 +
 +
=== Better IO with ranges ===
 +
 +
Ideally `std.stdio` should have a better range support. So far
 +
[https://github.com/jasonwhite/io io], [https://github.com/schveiguy/iopipe iopipe] and [https://github.com/rejectedsoftware/vibe.d/blob/master/stream/vibe/stream/stdio.d Vibe.d's streaming interface] emerged, but haven't got sufficient testing.
 +
`std.io` is reserved for the new io module.
  
 
=== Eventloop ===
 
=== Eventloop ===
Line 72: Line 78:
 
This new package could include a comprehensive set of encoder, covering (in addition to base64 and utf8) all the baseXX encoders,
 
This new package could include a comprehensive set of encoder, covering (in addition to base64 and utf8) all the baseXX encoders,
 
since they are easy addition (base16, base32, z85) and also domain specific encoders such as percent encoding (which then could be used in any replacement of the old-fashioned and incomplete ''std.uri'' module) or variable length quantity. Obviously the final list of encoders should be discussed on the NG.
 
since they are easy addition (base16, base32, z85) and also domain specific encoders such as percent encoding (which then could be used in any replacement of the old-fashioned and incomplete ''std.uri'' module) or variable length quantity. Obviously the final list of encoders should be discussed on the NG.
 +
 +
=== std.decimal ===
 +
 +
A general decimal math library for fixed point arithmetics.
 +
Existing libraries are [https://github.com/andersonpd/eris/ eris] and [https://github.com/jaypha/fixed fixed].
 +
 +
=== Other questionable modules ===
 +
 +
This list is a brief overview of developed replacements in the community.
 +
It is intended to show which modules might deserve a bit more attendance:
 +
 +
- std.base64
 +
- https://code.dlang.org/packages/base-d (lazily)
 +
- https://github.com/e10s/d-base32 (base32)
 +
- std.bigint vs. [https://github.com/andersonpd/eris/blob/master/integer/extended.d extended]
 +
- std.csv vs. [https://github.com/eBay/tsv-utils-dlang tsv-utils] (in general a CSV module should have a [pandas.pydata.org/pandas-docs/stable/10min.html Pandas]-like API with full range support
 +
- std.getopt - there are many community solutions:
 +
- https://github.com/jasonwhite/darg
 +
- https://github.com/SirTony/commando
 +
- std.math vs. [https://github.com/libmir/mir-math mir-math] and [https://code.dlang.org/packages/ctstdmath ctstdmath]
 +
- std.net.isemail vs [https://github.com/anton-dutov/mail mail] and [http://vibed.org/api/vibe.mail.smtp/Mail vibe.mail.smtp]
 +
- std.signals vs. [https://code.dlang.org/packages/phobosx PhobosX.signals], [https://code.dlang.org/packages/observe observe]
 +
- std.socket vs. [http://vibed.org/api/vibe.http.websockets Vibe's sockets
 +
- std.uri vs [https://github.com/rikkimax/alphaPhobos/blob/master/source/std/experimental/uri.d alphaPhobos], [http://vibed.org/api/vibe.inet.url/ vibe.inet.url], [https://github.com/adamdruppe/arsd/blob/master/http2.d arsd]
 +
- std.variant vs.[https://github.com/s-ludwig/taggedalgebraic taggedalgebraic]
  
 
== Walter and Andrei's Action List ==
 
== Walter and Andrei's Action List ==

Revision as of 19:40, 20 December 2016

Wish list

Containers

No one is working on this. It's tough to get started because, despite lots of discussion at various times on this forum, no one seems to really know what they want. Since the containers in question are well-known, it's much more a design problem than an implementation problem.

Update: there has been a lot of progress made towards GC-free containers with EMSI Containers.

Compression/archiving

Opening standard compressed/archived file formats needs to just work. This includes at least zip, gzip, tar and bzip2. Of course, zip already is available and gzip is supported by the zlib module but with a crufty C API. At least gzip and bzip2, which are stream-based as opposed to file-based, should be handled via streams, which means that streams block compression/archiving. Also, since tar and zip are both file based, they should probably be handled by the same API, which might mean deprecating std.zip and rewriting it.

Encryption and hashing

This is more an implementation problem than a design problem. No one is working on it. Some work has been done here but it's unfinished. One of the ideas is to wrap OpenSSL? at first and then implement the most useful crypto primitives in D to avoid library dependency and to make them usable with CTFE.

An improved std.json

Nobody is working on this right now. Wish_list/std.json documents the issues with the current std.json and some requirements for a replacement. Update: See std.data.json for recent development.

An improved std.xml

Maybe Thomas Sowinski is working on a replacement, but haven't seen any updates in a long time. There was also xml2 which was abandoned too.

Update: Lodovico Giaretta is working on this as part of this GSoC project - WiP code

Matrices and linear algebra

Cristi Cobzarenco's GSoC? project is a good starting point but it needs polish. He has indicated that he wants to get back to working on it but doesn't have time.

Update: Matrices are now merged part of Phobos (std.experimental.ndslice) and Ilya Yaroshenko started a new project - mir which will cover more numeric functionalities.

std.database

This is a large, complicated project because we're trying to define a common API for a variety of RDBMSs. Again, it's more a design problem than an implementation problem. There is at least one current attempt at a unified interface: DDBC (See also: [Database_Libraries]).

Parallel algorithms

dsimcha has implemented a decent amount of these in his std.parallel_algorithm Github project, but has become somewhat frustrated and unmotivated to finish this project because so many of the relevant algorithms seem memory bandwidth bound and aren't substantially faster when parallelized than when run serially.

Native curl replacement

std.net.curl is a wrapper around the low-level bindings to the C library curl which has to be included in the D release. Apart from the licensing issues, it also means that maintaining and extending std.net.curl is quite hard. A replacement library is already WiP, but still needs help and tester to get ready for primetime.

Native SQLite replacement

etc.c.sqlite3 is a wrapper around the low-level bindings to the C library SQLite3 which has to be included in the D release. Apart from the licensing issues, it also means that maintaining and extending this library is quite hard. There's already a SQLite Reader in alpha that takes CTFE optimizations into account, but of-course write access needs to be provided too.

Native ZLib replacement

etc.c.zlib is a wrapper around the low-level bindings to the C library Zlib which has to be included in the D release. Apart from the licensing issues, it also means that maintaining and extending this library is quite hard. The best existing library is archive.

Better IO with ranges

Ideally `std.stdio` should have a better range support. So far io, iopipe and Vibe.d's streaming interface emerged, but haven't got sufficient testing. `std.io` is reserved for the new io module.

Eventloop

Event loops have three common types.

1. Audio

A real time loop to play music (usually only concerned with one thread).

2. Windowing

Per thread event loop that must take input from the OS and move events to the appropriate handler. May utilise multiple system event loops per instance, such as X11 and Wayland.

3. Sockets

Per application event loop that takes ownership of events and messaging to threads.

Threading is also usually included with sockets for the usage by multithreaded web servers. A D (and in Phobos) implementation should take this into consideration with support of non-D event loops.

At the moment D's flagship web framework Vibe.d depends upon libevent - a C eventloop library. Installing this library often creates troubles as well as optimizing it. D should have a proper cross-platform event loop library with multithreading support that can be used as a general-purpose asynchronous library. Libasync is work in progress, however more testers and help is needed to make it a rock-solid library.

std.encoding

std.encoding needs a complete overhaul. Other encoding packages as ascii, base64, utf8 could be included and a common API should unite them.

This new package could include a comprehensive set of encoder, covering (in addition to base64 and utf8) all the baseXX encoders, since they are easy addition (base16, base32, z85) and also domain specific encoders such as percent encoding (which then could be used in any replacement of the old-fashioned and incomplete std.uri module) or variable length quantity. Obviously the final list of encoders should be discussed on the NG.

std.decimal

A general decimal math library for fixed point arithmetics. Existing libraries are eris and fixed.

Other questionable modules

This list is a brief overview of developed replacements in the community. It is intended to show which modules might deserve a bit more attendance:

- std.base64

- https://code.dlang.org/packages/base-d (lazily)
- https://github.com/e10s/d-base32 (base32)

- std.bigint vs. extended - std.csv vs. tsv-utils (in general a CSV module should have a [pandas.pydata.org/pandas-docs/stable/10min.html Pandas]-like API with full range support - std.getopt - there are many community solutions:

- https://github.com/jasonwhite/darg
- https://github.com/SirTony/commando

- std.math vs. mir-math and ctstdmath - std.net.isemail vs mail and vibe.mail.smtp - std.signals vs. PhobosX.signals, observe - std.socket vs. [http://vibed.org/api/vibe.http.websockets Vibe's sockets - std.uri vs alphaPhobos, vibe.inet.url, arsd - std.variant vs.taggedalgebraic

Walter and Andrei's Action List

Walter and Andrei have veto rights when it comes to inclusion of new stuff into DMD, druntime, and phobos. So if you're looking to work on new stuff either pick something from the list. Or get in contact with Walter or Andrei early during the design phase of your project. This way you can avoid a lot of frustration.

Walter's List

discussion

Id Description People working on
1 eliminate all gratuitous use of gc
2 review all Phobos modules for compatibility with ranges - std.zip, for example, was done before ranges and does not work with them
3 use -cov to improve code coverage of Phobos modules Browse annotated files on Codecov
4 make sure every function in Phobos has an example
5 make sure every function in Phobos has Params: and Returns: sections http://www.digitalmars.com/d/archives/digitalmars/D/Phobos_Documentation_-_call_to_action_258777.html
6 replace std.xml with something we can be proud of https://github.com/lodo1995/experimental.xml
7 compile a list of popular modules in other languages and see what we should have
8 create a greenthreads module that works like Goroutines
9 create a module that enables code to be run on GPUs https://github.com/thewilsonator/ldc/tree/dcompute
10 now that D can interface to C++ templates and exceptions, create the interface code to the C++ STL
11 review all of Phobos for @safe compatibility
12 remove dependency on autodecode from Phobos
13 there was recently a thread on Phobos desperately needing a BigDecimal module, someone helpfully posted links to a couple of free C versions, I suggested that someone build a D wrapper for one or convert one to D. Sadly, that was the end of the discussion
14 I regularly ask people to write articles about D, and the situation is slowly getting better, in that more and more good ones are getting written

Andrei's List

Id Description People working on
? Complete implementation of @safe Discussion at: https://forum.dlang.org/post/auowltizusrdhzjbimlg@forum.dlang.org