pouët.net

I'm going to add source links to zip files in addition to the existing github links on my prods. Is that okay?

yeah that's fine

Perhaps we can at least remove “googlecode” as an example in the FAQ, given that there is not a single “googlecode” extra link in any prod?

Speaking of GitHub, I noticed that there are some links involving GitHub violating the “link directly to the file, not to a download page”. E.g. this prod, which links to https://github.com/42Bastian/lynx_hacking/blob/master/lniccc2000/lniccc2000.lnx. It needs a ?raw=true at the end to be a direct link. I think replacing /blob/ with /raw/ would work, too, but GitHub is weird. I don't know the exact pattern here, or how many there are.

hmm, i wasn't aware googlecode was that dead, pretty sure there was a bunch of links last time i checked (which is addmittedly years and years ago). will have a peek soon(ish) and remove it from the FAQ if there's really nothing

as for main downloads on github or other sites with some sort of bufferpage other than scene.org, we're trying to root those out but it's a long and winding road

Prod 86469 has a “google play” link, that's the closest I can find :-) And 58283 / 58835 have links to code.google.com that are not marked with a “googlecode” type.

I still don't understand why we can't simply have a "repository" or something that can point anywhere? Why does it need to be popular? Why can we have dodgysite.com as download link but not commercialsite.com as a source repository? Whst purpose does this serve?

Please explain this like I'm five and I'll shut up. Github is terrible.

"use the big corporate platform or don't get linked" doesn't feel very demoscene to me

what preacher and yx said

Since I self-host my own git service, I guess I'll have to link my sources as "minisite" instead of "source" then... Thanks for letting me know, I guess.

petite pute de luxe

I think this is the wrong way around.

I'm the one who makes my productions, and I'm the one who decides where I host the code. If pouët doesn't want to link to that, that's pouët's loss. There's always demozoo.

If pouët admins want's to let these arbitrary nonsense rules allow the site to fade further into obsolescence, so be it.

and ra said to alesc/nooon :

Demozoo ? c'est quoi ? moi je m'en branle de demozoo .

There's a number of prods where the download link 404s, but cardboard doesn't understand it (usually because the HTTP response actually comes back with 200 OK but then some random HTML instead). A random example is this prod, I think there's maybe 50 of them in all.

With extra links, I can just send removal requests (at least in the common case of “video, and there's already a working youtube link”), but with prods, we can't remove it and changing to “http://127.0.0.1/no-download” or whatever removes important information about the file name and such. :-) Is there a way to feed this kind of data back into cardboard, so that glöperators will actually be on the lookout?

Similarly, there's a bunch of URLs that cardboard says are broken (usually with HTTP status 0) that are actually OK.

(PS: If people add Codeberg support, my mirror now clones git repositories and stores them as a packfile/bundle.)

in french please ...

i'm keeping a list of all prod ids that come up via cardboard when i do systematic checks at semi-regular intervals, and that example is in that list somehow, so it must have come up in cardboard at some point. would be interesting to see the rest of that list, and also the false positives, so if you can send them over that'd be much appreciated indeed.

havoc: link me beautifull

I haven't checked that all of these actually show up as good in cardboard, because I don't have programmatic access to cardboard and it's sort of annoying to scrape. But I've done spot checks.

(This also includes only the very simplest cases, namely the URLs ending in .zip that turned out to be text or HTML somehow)

And here are the files from the “http 0” cardboard page that both my mirror and my browser manage to download just fine:

ok cool, was able to replace most of those with better links. the files hosted on untergrund.net don't actually download here though ("file wasn't available on site"). maybe we should set smth up to transfer those files to a better host.


checked a couple handfulls of these and it's like 50% known and 50% new cases, so pretty useful overall. will check more deeply when i find some more time

I have a list of files that are also on scene.org (same hash except for scene.org.txt in zip files), if you're interested. It's ~1700 URLs, IIRC.

yeah sure, that might take a while to work through but better too much than too little data like that ;) is the 256b.untergrund.net stuff on that list too? cause that seems to be the bulk of the stuff i can't seem to grab which you could

My assumption was just that you could run a query and just change all of them. :-) I can file automated change requests but I'm not sure if you want to click through all of those (and I'd assume it'd count as abuse). Especially since I've seemingly filed so many now that some of them were forgotten (a couple of the older ones still sit as pending).


There are 40 256bytes.untergrund.net-to-scene.org matches in that file. 98 further files from 256bytes.untergrund.net (that are linked from Pouët prods) appear to have no match on scene.org.

I'll send you a copy of the list.

That's possible but not strictly necessary for stuff that's on other trusted servers, I see such a list more as useful backup in case things might go wrong :)


Thanks! :)

? ? ? ? ?