I might start on a project I've had in mind for a few weeks: a program that takes a DOI, ISBN, title or something related as an argument, finds the ebook (epub, pdf, djvu, etc), and downloads it. So instead of checking libgen, several trackers, and ebooks IRC channels, I'll just run the program once and it'll fetch the book for me.

I'm not sure whether I'll write it in Python, C++ (without Boost) or both (just for learnings sake), but I'll be a good alternative to the IRC library I'm working on; I've never written a serious one-shot program before. I think I'm calling it "bookwyrm".

(12-09-2016, 07:28 PM)iomonk Wrote: I'm thinking about getting started on my file tagging project.

I stumbled upon TMSU today but I have yet to try it. Do you mean something like that?

(26-09-2016, 03:53 PM)Tmplt Wrote: I might start on a project I've had in mind for a few weeks: a program that takes a DOI, ISBN, title or something related as an argument, finds the ebook (epub, pdf, djvu, etc), and downloads it. So instead of checking libgen, several trackers, and ebooks IRC channels, I'll just run the program once and it'll fetch the book for me.

Something like this would have been handy last semester for that paper, scavenging for things based on DOI. You should totally go on with this.

(26-09-2016, 04:04 PM)rocx Wrote: You should totally go on with this.

Good thing there is a want for this. I expected a reply pointing to an already existing project.
I have a lot of free time this semester. I'll probably start working on it this week unless something comes up.

So i finally fixed that weird bug in my touch-command program, whatever it is called. I also have a video for you guys. Unfortunately, the image seems to be too compressed or something, so the image will look a bit blurry, but you should be able to understand whats going on.

Anyone have an idea where I can upload videos besides youtube, so i can give a link?

(29-09-2016, 02:19 PM)mrmister Wrote: Anyone have an idea where I can upload videos besides youtube, so i can give a link?

There's https://teknik.io/

I've been learning some Rust, but it seems to change a lot of old concepts i had about programming, so it may take me more effort than expected.

And heres the demo video. Thanks for the help albertocg.

https://u.teknik.io/4UNEv.mp4

EDIT: Looks like the file is actually corrupt. Will try to give another try

(29-09-2016, 06:31 PM)mrmister Wrote: And heres the demo video. Thanks for the help albertocg.

https://u.teknik.io/4UNEv.mp4

EDIT: Looks like the file is actually corrupt. Will try to give another try

No, it's not corrupt at all, I can see it! It's just teknik extremely slow. Great job with your project!

Sometimes I port other people's applications to several platforms instead of writing my own ones. These days I finally managed to get my first Chocolatey package approved, it's the TECO editor. Hooray!

blogcpp v3 is released.

Oh, cool.
Which C++ do you use? 98/03/11/14/17?

Mostly C++17 (using the std::filesystem library) with some C.

I've begun my work on bookwyrm but before I get to the serious work on the innards, I think knowing the how the (command line) interface should be will help a lot. If you have a program alike how I described it here, how would you like to specify arguments? Good old


or something like beets:


or just combining all arguments into a single search query?

I had a thought to take very few arguments from the user, search for a book or paper on some database(ses) and get the rest of the data from there before starting the search for a download URI. But perhaps that is overdoing it?

I'd definitely prefer POSIX style (-t -a...).

Yeah, I agree with jkl.

It is easier for the eyes, it fits well into every unix system,
and it is what people expect.

Most programs I make that need command line arguments are normally Perl scripts, so I use Getopt::Long, which lets you specify long names (e.g. '--title="foo"') which then automatically allows short arguments too (e.g. -t "foo"), so I've come to appreciate this style.

I guess my preference is shaped by arg.h, which uses plan9 style
macros for argument parsing and does not allow anything
besides -t foo.

Thanks for the input. I'm going with POSIX-style.

First: it is still some half-baked draft.

I am learning a few C. I am trying to do some dmenu-style tool for the terminal. It is really not the first time that someone try:

- fzf (go): https://github.com/junegunn/fzf
- fzy (C, that I use): https://github.com/jhawthorn/fzy
- peax (awk): https://github.com/josuah/Config/blob/87...e/bin/peax
- vis-menu (C, fork of slmenu): https://github.com/martanne/vis/blob/master/vis-menu.c
- slmenu (C, fork from dmenu): https://bitbucket.org/rafaelgg/slmenu/
- pick (C, from thoughbot): https://github.com/thoughtbot/pick
- pick (C, from calleerlandsson): https://github.com/calleerlandsson/pick
- selecta (ruby): https://github.com/garybernhardt/selecta
- ivy (emacs lisp, emacs-only): https://github.com/abo-abo/swiper

And a lot more, in a lot more languages...

This one will not be fuzzy, but in the future, spaces will be treated as wildchars.

It purpose is to get the whole vimperator completion workflow:

1. Start to type to match a command and get grayed completion after the input;
2. Tab to complete;
3. Space to select the command and start to type the next;
4. Have comments at the end of the lines, describing what each command does;
5. Different sections (I call them headers) from different kind of completions, all at the same time;

Any other command can not have 'space' as a key to validate a command.
And none can have headers or comments.

I also want to bring the Swiper workflow from emacs: have the lines from vim buffer filtered out in a completion engine like this one, and each time I move up or down in it, jump to the corresponding line in the vim window accordingly.
https://github.com/abo-abo/swiper

I only spent a day writing it, so I am far from finishing it!

If curious, maybe you can compile it and get a segfault at http://github.com/josuah/iomenu

I'm currently writing an audio player in Delphi just for the sake of it. Oh, and I just added configurable permalinks to blogcpp (because I've read a blog post asking for a blog generator which has them). :-)

(17-10-2016, 07:09 PM)jkl Wrote: I'm currently writing an audio player in Delphi just for the sake of it. Oh, and I just added configurable permalinks to blogcpp (because I've read a blog post asking for a blog generator which has them). :-)

I did not know about Delphi! Yet another programming language discovered on nixers!

Delphi is one of the Elder Gods. :-) (More or less.) While it seems to have been stagnating since Borland decided to focus on existing Enterprise customers (and many "old" Delphi users switched to C# or similar weirdnesses), the birth of Lazarus (a Delphi clone based on Free Pascal, almost compatible with the Delphi syntax) has given it a new life. And it's still awesome: Its ALGOL roots gave it great features like a clean syntax and a damn fast compiler; plus, recently, it got support for developing cross-platform apps, including mobile.

Pascal/Delphi was my second "real" programming language after Visual Basic. It's nice to play with it again after all these years.

It looks to be one of those precious relics from the past, with more recent initiatives to freshen it up, like UNIX.

Atm, I am learning lua, klong (www.t3x.org/klong) and hope (https://github.com/shaos/hopeless). Also, I have been reading a bit about neural networks, as well as working quietly on some semi-secret projects (they wil take some time to be usable, that's why secret).

(18-10-2016, 04:40 AM)pranomostro Wrote: Atm, I am learning lua, klong (www.t3x.org/klong) and hope (https://github.com/shaos/hopeless). Also, I have been reading a bit about neural networks, as well as working quietly on some semi-secret projects (they wil take some time to be usable, that's why secret).

Small, functional, programming languages, with implementations in C: Sounds good so far!

It would take me some time to get used to Klong notation, though.

(18-10-2016, 02:10 PM)josuah Wrote: Small, functional, programming languages, with implementations in C: Sounds good so far!

Yeah, they are quite small implementations. For klong, I am solving the 99 lisp problems
(http://www.ic.unicamp.br/%7Emeidanis/cou...blems.html),
which works quite nicely with array languages.

(18-10-2016, 02:10 PM)josuah Wrote: It would take me some time to get used to Klong notation, though.

It's actually not that hard. And array languages have a right approach operators, they just
like them too much:


Solution for problem 26, subsets of size y in x.

klong and hope both lack utf8 support, unfortunately. But that could be solved.

Currently i'm learning kotlin. Slow and steady. After trying scala and groovy, i think i've found my favourite jvm language.

At work we're currently stuck with java 7, so i really need something more concise in the evenings to keep me sane.

My favorite JVM language would be one which doesn't need the JVM to work.

Interestingly, I can't delete a post, instead, it leads me to a page 'You don't have access to this page'. Take this as a deletion of what I wrote before.

(18-10-2016, 04:40 AM)pranomostro Wrote: Atm, I am learning lua, klong (www.t3x.org/klong) and hope (https://github.com/shaos/hopeless). Also, I have been reading a bit about neural networks, as well as working quietly on some semi-secret projects (they wil take some time to be usable, that's why secret).

Woah, Klong looks great! I've tried a couple of times to break into a compact array language like J or K, but I've been struggling. Somehow Klong seems a little easier to understand. Perhaps that prime example is just really well written!

(20-10-2016, 09:45 AM)lemons Wrote: Woah, Klong looks great!

Wonderful. I had the same feeling, I started learning 2 years ago but then gave up because the language is so big.
Klong is for some reason comparatively accessible.