<ijp>if you remove them, boot-9 "only" takes 1.3MB <ijp>I conjecture there are similar savings in a lot of files <catonano>spk121: I neve had that but I'm happy about thhis fix anyway <Muto>I built Guile from source on Debian stable, but now whenever I 'apt upgrade', at the end of the list I get a "libguile.so is not an ELF file: it has the wrong magic bytes at the start <daviid>Muto: you can safely ignore this warning <sneek>Welcome back daviid, you have 1 message. <Muto>daviid: Okay, thank you. I was just worried in case it was something bad. <daviid>Muto: no, nothing to worry about, though I don't recall the detail, someone else may explain if needed, but it's a 'false positive' warning <daviid>paroneayea: yes, I did see that message, tx. I'm very much in favor of slowly but surely grow Guile-Lib with Guix 'stable' and well identified functionality that might precisely 'land' in Guile-Lib, just as Ludovic answered, if not in Guile core directly... But I couldn't find the time to address this, yet have to answer Alex and Matt for their offer to contribute (plist, regexp ...) <daviid>paroneayea: note that guile-lib is a bit of a 'strange beast' as it is :), because the manual is generated form the source (which is not the way I usually work and I don't know the 'details', cause so far I mererly contributed on the admin side, to keep guile-lib 'alive', modernize its autotools ...), and the tests use its test-suite :), based on goops, which is perfectly fine for me, but guix uses srfi-64 if not mistaken... <daviid>paroneayea: but nothing two difficult either and Guile-Lib really need to become what it always ment to be: "an accumulation place for pure-scheme Guile modules, allowing for people to cooperate integrating their generic Guile modules into a coherent library. Think "a down-scaled, limited-scope CPAN for Guile"." <daviid>paroneayea: so, help welcome of course! let's try to get Alex's plist in Guile-Lib as a little exercise ;), wdyt? Or some guix functionality if you prefer, I guess what I mean to say is I'm ready to play with the idea but I need help (time wise principaly, but module names, doc, tests and design 'wise as well ...) so everything we do is 'robust', security 'free', well tested, debuged ... <daviid>paroneayea: I did think about changing the web site as well, nothing about the previous design of course, but to get a 'fresh wind' blowing on it :), help welcome there as well... ***random-nick is now known as Guest33714
***random-nickname is now known as random-nick
<reepca>Is there a function equivalent of AND and OR? trying to pass them as functions causes complaints about failing to match syntax patterns <civodul>sometimes 'every' and 'any' from SRFI-1 are what you want <sporgj>what exciting stuff is there for a newbie? <cmaloney>sporgj: How familiar with scheme are you? <sporgj>Over the summer, I wanted to get familiar with scheme/functional languages <sporgj>tried haskell and others: didn't like their toolchain <sporgj>guile just seemed the easier to install and maintain <sporgj>cmaloney: I'm current going of the SICP <sporgj>OrangeShark: my hope is to become a developer for the language and standard library <sporgj>OrangeShark: I'm thinking mostly about improving things like the REPL interface <cmaloney>I'm in a similar boat (wanting to get my feet wet with functional languages and not have that introduction be 100% through JavaSCript) <sporgj>and to me it's the best scheme I've played around with <cmaloney>Yeah, Guile (And Racket) have tickled my fancy. <OrangeShark>sporgj: something you might want to check out that was discussed yesterday is guile-lib, daviid mentioned wanting to revive the project and add more modules. <sporgj>cmaloney: lol I think you've spying on me <sporgj>cmaloney: I installed Racket too <catonano>cmaloney: you know about the SICP, don't you ? <sporgj>has a lot of libraries but the toolchain is not the best <sporgj>catonano: yes that's what I've been going through <sporgj>while applying guile-specific code <cmaloney>catonano: I've been rading SICP and Realm of Racket <catonano>sporgj: very good. When I went trough the SICP I used the MIT-scheme because I didn't want to run into ANY incompatibility <cmaloney>I've had a hard time with the style of that book <catonano>cmaloney: ok, I never trried the little schemer, I know it's famous <catonano>cmaloney: I tried with a book on relational programming. The style is liek the one of the little schemer. I have to say I don't lie it either <cmaloney>It is. I'm sure it works for some but it didn't work for me. <catonano>the SICP is great but I agree that a less intense introduction to scheme would be useful <sporgj>catonano: please refer me any resources that could be of use <cmaloney>same. I found common lisp a little too dense for me <catonano>sporgj: well, if you start from Clouure there are tons of materials, bothh books and footages <catonano>there is "Cloure for te brave and true" you can read it or free on line <catonano>thre are also some classes as video footages <catonano>Clojuure is a lisp but it's a way to get started <cmaloney>(played with Land of Lisp, which is excellent, and a litle of Clojure for the Brave and True) <catonano>you can use Emacs, it has a great environment for Emacs <cmaloney>(confession: I used to use Emacs and now I can't handle it since I got comfortable with Vim) <catonano>For example I was introduced to te game of life in clojure. Ten I made the same in clouurescript, it ran in the browser. I saw a game of life made with Guile and the Gnome native stuff <cmaloney>Yes, and I kept finding the 20% that wasn't implemented. <catonano>well clojure has some great support for Vim too <catonano>and I guess that movin Vim to Guile shouldn't be that terrible <cmaloney>sporgj: I've found that SICP is a little bit dense in the comp sci theory <cmaloney>but it's helpful because you're trying to learn new ways of doing things <sporgj>cmaloney: I'm a grad student, so it's not so bad <sporgj>jus needed the exposure to this form of thinking <cmaloney>Ah, it's been 20 years since I was in a comp sci program <sporgj>but will definitely look up with clojure stuff <cmaloney>So yeah, that will likely be an easier read for you now <cmaloney>Have you looked at the guile manual proper? <cmaloney>The quick intro to scheme is well-written <sporgj>cmaloney: yeah, for things like how to read command line arguments <cmaloney>ACTION tries to think of resources that aren't Racket. <sporgj>cmaloney: Racket has a lot of stuff <sporgj>it's like the perfect midpoint between Latex & Org-mode <cmaloney>I've been working through a little bit of Scheme / Racket on my blog <catonano>I gotta go. I'll read yo again soon ! Bye ! <impossibru>Is it possible to save guile's execution state somewhere to a file, shutdown guile, and resume it with the saved state in the file? <cmaloney>impossibru: Not sure what the internal support in guile is for that but the OS may support it <cmaloney>are you looking for something like Python pickling? <cmaloney>pickling takes an object and writes it to disk <impossibru>I was thinking of using Guile in an Android app, and at any point the OS can kill the process. I think other than a callback that is called before the event happens, there are not many builtins to save the state of running processes (but I'm not 100% sure) <dsmith-work>impossibru: Pretty sure nothing like that exists right now. Guile *used* to have an unexec a long time ago. <impossibru>cmaloney, I think that would help if it also works for the evaluator itself <davexunit>saving the running state of a process is a tricky thing <davexunit>the only thing that really makes sense is to write a custom procedure that will serialize your application's state to an s-expression and write it to disk. <impossibru>Is it possible to get the state of Guile itself? <impossibru>I'm thinking of the case where Android kills Guile while Guile is eval-ing something. <impossibru>It might be possible to suspend termination until after the eval is done, though... <davexunit>what happens in the event of a software upgrade? <davexunit>seems like the wrong level of abstraction to be working in <avoine>wingo: in fibers, would it make sense to add an (all-operation ops) to synchronize more then 2 channels or it needs changes in the way channels work? <avoine>or more likely I'm doing it the wrong way ;) <amz3>I was failing to see the obvious, what wiredtiger needs is not a shiny new toy DSL, but a good old standard shit like SQL <amz3>I say the above like that because in my mind, I wondering whether it's more useful to improve the DSL to match better SQL or to create an end user app. <amz3>I've been basically feedling with no-sql stuff for 5 years <amz3>going back with better knowledge of databases and looking at SQL allows me to say that SQL is well suited for low complexity queries <amz3>it does midly complex queries very well <amz3>it factors common query patterns <amz3>the thing in SQL I'd like to have is the ability to simply 'drop' to a real language <amz3>for instance, it happens that I am doing two loops in serial to correctly render a html report, where a single loop could be enough <amz3>a language that I already know I mean <amz3>prolly using pl/sql in postgresql you can do whatever you want, but I don't know pl/sql <amz3>also when using pl/sql your are half forgeting about vendor-lockin <amz3>I will add SQL in wiredtiger on my todo list <amz3>OrangeShark: yeah, Linq is similar to traversi framework in grf3 library <amz3>I never used it for the real, but most ORMs have inherited the "Linq-way" of querying data <amz3>it was like a revolution when it came out <OrangeShark>you can either do this sort of object based way or write SQL like code in C# <amz3>that's a great idea, I would love that <amz3>that's awesome, but... they have several backends for the API or they have an abstraction on top of the rdbms table to threat it like collection? <OrangeShark>amz3: I mean they have the same API over their collections <OrangeShark>so treat both databases and collections with the same code <amz3>I am wondering how they bridge the gap between databases and simple collection, because the API of databases is much different compared to Collections' API <ecraven>amz3: is it, though? things like filter, find, map, with accessors, would work well on both, I think? <amz3>ecraven: 'slice' is missing from that list <OrangeShark>amz3: looking at the documentation, seems it used an interface for the collections and other data sources <ecraven>yea, but I think a collection of objects and a traditional relational table are not that different, with a bit of squinting ;) <amz3>the problem I see is that if you threat a database like a collection you won't be able to take advantage of performance tricks <amz3>OrangeShark: that's what I was reading <OrangeShark>amz3: I think it is abstract enough that it can perform some optimizations for performance. I haven't really looked too much into detail of how it works. <ecraven>amz3: that might depend on your query language <ecraven>and of course your use case.. if you have Big Data and want to work with Rollup and Cube and friends, you might want to go to dedicated solutions <OrangeShark>so whatever was built up can be manipulated before you query the database <amz3>ecraven: what is Cube and Rollup? <manumanumanu>has anyone made something similar to rackets for-loops in guile? <stis>I have a direct port of almost all that <stis>otherwise I think that the standard reference is foof loop <manumanumanu>I am half-way there using syntax-case. no error reporting whatsoever, but it is a fun exercise <stis>look for compat/racket/for.scm <manumanumanu>stis: does it handle internal loops using #:when/#:unless? <stis>manumanumanu: no it does not handle #:when and yes syntax parse is mighty cool <Muto>Has anyone managed to get guile-reader built from tarball with Guile-2.2? <manumanumanu>stis: is it anything like the racket one? or do I have to read source code to get any documentation? <stis>it's a port of racket's syntax-parse since say 3 years ago, standard stuff should be the same <stis>you can also ask here if i'm available <stis>so please use racket docs <Muto>I can't build guile-reader on Debian stable with Guile 2.2. Is there a way to do this? <stis>I just saw that I had a racket for not checked into a repo, I just did a repo guile-for for it <stis>this is the second generation <ijp>js proxies are so helpful for debugging <ijp>weird, it looks like the vm primitive for apply takes one more argument than it used to <ijp>that can't possibly be right, since it would show up in blame