Posts by carlozancanaro@mastodon.technology
(DIR) Post #538809 by carlozancanaro@mastodon.technology
2018-10-14T21:02:03Z
0 likes, 0 repeats
@cwebber Monday morning here, so perfectly timed!
(DIR) Post #2181186 by carlozancanaro@mastodon.technology
2018-12-22T03:00:39Z
0 likes, 0 repeats
@technomancy This is harder when we need to interact with other systems, built by other people. How do you make IM software that "chafe[s] at structures"? The software needs to interoperate with others, and thus can't be purely personal. Hence, standardisation.I'm really on board for people having the ability to build personal systems, but to do that in the modern world I think we need standardised "languages" to communicate with other peoples' software (whether personal or corporate).
(DIR) Post #2638357 by carlozancanaro@mastodon.technology
2019-01-04T11:21:50Z
0 likes, 0 repeats
@technomancy Are you still using EXWM? AwesomeWM is being a pain for me (ie. I'm tired of carrying a patch for [1] in my local Guix repo, and there hasn't been a release in ages), so I'm thinking of trying out EXWM and/or StumpWM.My biggest fear with EXWM is that I'll try to open a huge file, lock up Emacs, and break everything. Does that happen to you?[1]: https://github.com/awesomeWM/awesome/issues/1344
(DIR) Post #3221015 by carlozancanaro@mastodon.technology
2019-01-20T20:37:38Z
0 likes, 0 repeats
@cwebber @emacsen Hey, in the Libre Lounge episode about video games I feel like I remember you mentioning a game that was looking for help with its soundtrack, but I can't remember which game it was. I have a friend who's interested in making video game soundtracks, so I'd like to point him towards it. Can you help me out? (I'll also accept an answer of "you imagined it", but I'll be a bit sad.)
(DIR) Post #3221017 by carlozancanaro@mastodon.technology
2019-01-20T23:40:10Z
0 likes, 0 repeats
@cwebber Thanks! I thought it was Empty Epsilon, but I can't find any reference to a soundtrack on their website. Maybe @emacsen can show me where he saw that?I had completely forgotten that the @librelounge existed. I definitely should have pinged that instead of you both. Oh well! Close enough!
(DIR) Post #3234371 by carlozancanaro@mastodon.technology
2019-01-21T23:39:01Z
0 likes, 0 repeats
@emacsen @librelounge Ah, okay. I must have misunderstood. I thought they were actively looking to improve their soundtrack.Oh well! Thanks for the reply, anyway!
(DIR) Post #9fmhYPrt16gjuG11No by carlozancanaro@mastodon.technology
2019-02-13T07:02:46Z
0 likes, 0 repeats
@emacsen @librelounge Ha! I was going to jokingly ask about that. Well done getting it under forty minutes!
(DIR) Post #9kmEYTqF3Yp4kvcPhI by carlozancanaro@mastodon.technology
2019-07-10T08:12:05Z
1 likes, 0 repeats
@cwebber Looking through what I can find, the Guile protocol is just sending forms, which is vulnerable to this attack because it ends up ignoring the GET and other extra bits.SLIME's protocol is s-expression based, and thus the GET will cause the message to be malformed (and, according to my tests, will cause swank on the CL side to terminate the connection).I'm not sure this is as serious a vulnerability as it was in Guile.
(DIR) Post #9w8VeHUVYb3h1QhMgK by carlozancanaro@mastodon.technology
2020-06-15T11:24:51Z
0 likes, 0 repeats
@jeko How do you feel about this, in general?I find multiple return values conceptually satisfying, but in practice they're just harder to work with than first-class containers of multiple values. As long as you have some method of destructuring (eg. see Python, Clojure, Javascript) or pattern matching (eg. see anything in the ML tradition) then I'm not entirely convinced that multiple return values are that useful. I guess they have performance benefits?
(DIR) Post #9xWh0Pg4BB7K6ZPr8K by carlozancanaro@mastodon.technology
2020-07-27T13:44:09Z
0 likes, 1 repeats
@ioa @phoe I think it's because a condition system really shines in an interactive environment (eg. like a Lisp REPL), whereas most languages are designed first-and-foremost to run as a static program. Condition systems really shine when they allow the user to interact with errors in the system and correct them on the fly, but that sort of interactivity doesn't really fit with the way most languages are conceptualised.
(DIR) Post #9xaKNfSg8jfFPC75gO by carlozancanaro@mastodon.technology
2020-07-29T08:39:07Z
0 likes, 1 repeats
@julm One thing this view is missing is that build reproducibility provides a mechanism for accountability.When there is an expectation that the binaries correspond exactly to the source, and there are tools to verify it, then vendors can be kept accountable when they violate that expectation.The goal of reproducible builds isn't to solve all potential supply chain attacks, but it provides an important foundation on which other things can be built.
(DIR) Post #A062OfvQIw3V4UGzzc by carlozancanaro@mastodon.technology
2020-10-12T20:17:45Z
0 likes, 0 repeats
@plausible "Their motives don’t seem to be to make the web more privacy-friendly ... . It seems purely a business opportunity to make money from open source."I understand this feeling, but expecting everyone to share your goals is unhelpful. It's good for companies to want to make money from open source. It's only a problem if they abuse it (not contributing, etc.).The license change is a good one because it sets a clear expectation, not because it stops companies making money.
(DIR) Post #A0651ZIJ5k0dhxeNFo by carlozancanaro@mastodon.technology
2020-10-12T20:24:22Z
0 likes, 0 repeats
@plausible I also think giving the sense that you chose the license *because* Google is against it (which your Toot does, even though the blog post doesn't) comes across as immature. I think the AGPL is a good license for what you're doing, but Google disliking it isn't a reason why it is good.
(DIR) Post #A06BKo0dCnxIZYwGv2 by carlozancanaro@mastodon.technology
2020-10-12T21:57:58Z
0 likes, 0 repeats
@plausible I agree. I just worry that statements like this:"Their motives don’t seem to be to make the web more privacy-friendly and reduce the dominance of Google. It seems purely a business opportunity to make money from open source." come across as "if you aren't completely on board with our philosophy, then we're not interested in working with you".
(DIR) Post #A3Ec20L6PnPUGRjbsG by carlozancanaro@mastodon.technology
2021-01-14T01:54:00Z
1 likes, 0 repeats
@scolobb Patches are welcome!I think there are a few reasons why manifests aren't emphasised.Manifest examples are harder to "follow along at home". Running commands is easy to follow in a tutorial. Creating a manifest file raises questions like "where should this file go?" which aren't hard to answer, but add complexity.Also, manifests are "foreign" to users who are familiar with apt/yum/pacman. Manifests are familiar in language-specific package managers, but that could be an unhelpful association.
(DIR) Post #A6nzguKZVtbsBAKHfk by carlozancanaro@mastodon.technology
2021-05-01T12:28:50Z
1 likes, 1 repeats
@veer66 If you're only using it as an alist, you can cons a new one on the front and it will shadow it for assoc purposes. Something like, (cons (cons key value) alist).If you're okay with mutation you can use (setf (cdr (assoc key alist)) new-value).Otherwise, I don't know. It feels like it *should* be easier than that, but I'm not familiar enough to know how.
(DIR) Post #A9wpUYthFbkwXafNYG by carlozancanaro@mastodon.technology
2021-07-19T03:06:39Z
0 likes, 0 repeats
@manyver_se Hey, are your F-Droid builds still working? My client thinks you haven't released since October last year, which seems at odds with your recent posts about releases.
(DIR) Post #AKO02gr8X2DwqvT1O4 by carlozancanaro@mastodon.technology
2022-06-11T04:30:31Z
1 likes, 1 repeats
@veer66 Due to the way alists work, you can actually just do (acons k v m) and lookups will find your new value instead of the old one. Depending on what exactly you're doing, that might be enough.You can also replace your remove -if call with (remove k m :test test :key #'car), which is a little bit simpler.