exists simple way for debuging/tracing macros?

@wwall I usually use the Macro Stepper in DrRacket

@notjack thanks. i understood this code

@pocmatos: from https://racket-news.com/2019/11/racket-news-issue-19.html:

> Some data about the activity in the Racket et al. repositories, for the month of September, 2019.

Shouldn’t this be October? As I understand, September is already featured at https://racket-news.com/2019/10/racket-news-issue-17.html

@sorawee oh dear, you’re right.

Thanks - will fix.

Should be fixed in the next 2 mins.

It looks like the package server has stopped picking up changes since at least yesterday. @jeapostrophe @tonyg

I added this package1 about ~24 hours ago and its metadata hasn’t been picked up yet. Other packages that I’ve changed 2–3 days ago haven’t been updated either.

This looks fixed now?

Looks like it!

Hey @jerome.martin.dev, and thank you! I did take a look at this collection and only did not use it due to a difference in coding style and audience. I intentionally leveraged a filesystem coupling so that changing the CLI could be very mechanical (e.g. rm removes a command, mv can put a command under a new parent, etc), and so that I didn’t have to keep editing a central specification. I definitely view my approach as niche, and useful only to those who think similarly about the problem.

Hey devs, I was about to cross post another project update here but I’m starting to feel like I’m contributing noise around messages for people asking for help. Would it be worth creating a Slack channel for people like me to share updates so I don’t distract?

Posting here is fine (it’s not that there are a big discussion going on at the moment).

Ok.

In that case:

New update on the Vulkan collection: https://sagegerard.com/racket-vulkan-contributors-update.html

Post it to Racket Stories too!

Where is that?

http://racket-stories.com\|racket-stories.com

Ah, sweet. Thank you

@soegaard2 Unfortunately I can’t register. Submitting the Create Account form results in a refresh with no feedback, even on the console. Trying to log in right after does the same. Using Firefox.

I’ll take a look.

Thanks. PM or email me if you need me to do some legwork.

I decided to be conservative (maybe too conservative) wrt user names.

@mflatt Re: _enum in a FFI, what’s the overhead on translating a symbol to the associated numerical value? Is it just a hash lookup?

Looks like it’s a association-list lookup. Maybe it should change to a hash lookup, especially if there are more than a few symbols.

Ah, thanks. I don’t know my worst-case, but there’s one example that comes to mind that has over 100 ~members~ enumerants.

Is there an inverse of cookie->header ? I need to handle session timeouts, and one option is for my “front controller”, which delegates to various handlers, to modify the expiration of a cookie. However, by the time the response is available, cookies have already been converted to headers. I will likely need to wrap request and response with my own data structures, but I was hoping to postpone that for a bit.

In hindsight, it seems that the response should have a set of cookies, and they would be converted to headers later in the pipeline.

Don’t know, but any time I see cookie and header processing that’s singular I get mildly terrified, because Set-Cookie is literally called out in the http header parsing spec as a special case that cannot be transmitted as a single header. For all other headers, you can merge headers with the same name by joining their values with ; (or maybe it’s ,, I don’t remember). But you can’t with Set-Cookie because it’s grammar was designed by browsers before there were any specs at all for http

I don’t really like the idea of parsing the header to get the cookie back anyway. I think I’ll need to create my own request/response data structures - best to do that when the number of handlers I have is relatively small.

:tada: https://github.com/zyrolasting/racket-vulkan is now at 1.0. Breaking changes will no longer be the norm during development.