patrick.net

You know I thought the link scrolled me down to the wrong comment in the thread, but I was mistaken - it just simply doesnt scroll to a comment at all. If the link doesnt give me the "invalid start number" error (most links give this error now), then it just doesnt scroll me down to any comments.

This link exhibits such behavior. It should take my to the last page where my comment is, and then also scroll me down to the comment, but doesnt. I see the first post in the thread, and then the first comment underneath it is 2116 out of 4485. No auto scrolling occurs.
https://patrick.net/post/1338647?start=2116#comment-1840406

I suppose this is because the url param start=2116 is wrong. The url should probably be start=4485 or something like that, to take me to the page where my comment is. I bet fixing the other bug fixes this too.

@Patrick - where is your current code again?

You need a regression test system. I've made a few. I need to look at (and setup) your code to see how I approach it. As I recall, you're using nginx, mysql, and your code. I can short circuit the email system so I can simulate that.

Curl should be able to simulate user interaction. Verifying what it produces should yield the same results.

Hircus says

1) When I click on my profile and then view my comments (https://patrick.net/comments?a=Hircus), they now start at my oldest comment, when it used to be my newest comment. I prefer the old behavior if this wasn't an intentional change.

@Hircus your comments should start with the newest page of comments now.

Hircus says

2) When viewing my comment list, if I click the date/permalink link of a comment, in the past it would take me right to my comment, but not anymore. Some links give me an "invalid start number" error, like this one
https://patrick.net/post/1345054?start=2118#comment-1840720

This is also fixed in that if a url has a start comment number which is larger than the number of comments on that post, it just shows you the last page of comments instead of quitting with an error message.

Hircus says

I suppose this is because the url param start=2116 is wrong. The url should probably be start=4485 or something like that, to take me to the page where my comment is. I bet fixing the other bug fixes this too.

Correct, the start param is wrong, because the page of comments which starts there does not contain the comment number in the hash tag.

Where did you get that url with the incompatible start and comment_id params? I should fix that if the url appeared on the site.

richwicks says

You need a regression test system.

I do have one, using mocha:

https://mochajs.org/?source=patrick.net

I have only basic tests for creating an account, logging in, creating a post and a comment, and a few other things like that.

I should create a few more, but:
1. I run them on every checkin to git, so I don't want them to take more than a few seconds.
2. I find that if my tests are too elaborate, I end up fixing tests all the time when I change the code, which is not productive.

Oops, clicking on quotes is broken because the start param is null.

Will fix.

Patrick says

Where did you get that url with the incompatible start and comment_id params? I should fix that if the url appeared on the site.

From https://patrick.net/comments?a=Hircus

Thanks, fixed that now.

Still need to fix the link in a quote.

OK, the link in a new quote should work now.

I still have to go back and change the database so that previous quote links work.

BTW, you can now search a person's comments on their comments page, eg:

https://patrick.net/comments?a=Patrick

@HunterTits

I removed the link to search for comments with images, because it is really just a search and you can search a person's comments now. So you can still see all your tit pics like this:

https://patrick.net/comments?a=HunterTits&s=width+height&start=3210#last

@HunterTits I made the search start on the last page of result if you include -1 as the start param, for example:

https://patrick.net/comments?a=HunterTits&s=width+height&start=-1#last

That way you can always get to your last comment with an image even as the number of comments changes.

If you want to see the whole last page of comments and not just the very last one, leave off the #last like this:

https://patrick.net/comments?a=HunterTits&s=width+height&start=-1

HunterTits says

It shows everything that has 'width' & 'height'..

Width and height, tits and ass, whatever.

@patrick bitchute videos dont have a link to the url like other videos do. example: https://patrick.net/post/1337984?start=324#comment-1842291

Thanks for telling me @Hircus

Should be fixed now, at least for input urls of the form https: //www.bit chute.com/embed/2xAdBjtpMRh1

(I put spaces it that so it would not show the video or create a clickable link.)

Please tell me about other things to fix. I'm on a roll lately, burning through my big list of patrick.net bugs.

@patrick this post was made today and didnt seem to show a link: https://patrick.net/post/1334038?start=239#comment-1844229

Thanks, dunno why.

I edited the comment and reposted the same bitchute link, and now it has an "original link" below it.

Wow, I can't register on patriots.win without the Google recaptcha spyware reporting back my details to Google/NSA.

To log in to Reddit also requires Google api spyware, so that's out too.

And I can't log in to Twitter because it demands a phone number, and that's a fuck no, never.

I'm kind of proud that none of that shit is on patrick.net, as anyone can verify.

How would I do that for free, aside from Google Voice?

Patrick says

How would I do that for free, aside from Google Voice?

there got to be a way. there are millions of fake accounts there somehow. although i do wonder if twitter employees create those via their systems bypassing all verification.

FortwayeAsFuckJoeBiden says

i do wonder if twitter employees create those via their systems bypassing all verification

I suspect that they do, either for their own personal reasons, or on CIA orders.

Not sure if someone requested. How about a tracking thread with newly uploaded at least non restricted images with a link to respective threads.

@Blue This is possible, but a bit complicated if you want images from both original posts (threads) and from comments.

It's pretty easy for just comments.

I guess link to original images should be good enough. It’s one quick way to browse and visit threads.

You can find all the latest threads which have images or videos in the original post like this:

https://patrick.net/search?order=new&s=width%20height

They are just thumbnail images though. And blank for videos.

I suppose I could make them bigger.

Blue says

I guess link to original images should be good enough. It’s one quick way to browse and visit threads.

@Blue Please tell me more about how you imagine this would look.

Would it just be a long list of images with links below them? Where would I put it on the site?

Thanks.

Patrick says

Would it just be a long list of images with links below them? Where would I put it on the site?

Yes. Create a thread something like, 'Picture2thread'. In case of s*x related images, I'd expect the thread name instead of the image, otherwise it might offend somebody.

Blue says

Patrick says

Would it just be a long list of images with links below them? Where would I put it on the site?

Yes. Create a thread something like, 'Picture2thread'. In case of s*x related images, I'd expect the thread name instead of the image, otherwise it might offend somebody.

@Blue OK, I did it.

Click the "latest images" links in the footer. That thread will fill up with images as people post them to an original thread or a comment thereon. There is an "origin" link below each image in that thread.

Here is that footer link for convenience:

https://patrick.net/post/1?start=-20

Let me know if I made any mistakes.

Looks good. Thanks Patrick for doing it.

I think I have added support for .webp images by converting them to .png images.

So if you've wondered why you uploaded a certain image in the past and it didn't show up, that was probably why. Now it should work.

Please let me know if I fucked anything up about image uploading.

Thanks @HunterTits

Can you give me an example url of an image ending in .txt?

HunterTits says

https://patrick.net/uploads/a/3/2/a325f7d2de5cc7c973c97eaa9fbf8c18.jpg

Can you point to a POST where a picture does this?

What OS are you using? I'm on a Linux system, I've never seen this happen, and there's plenty of pictures I've saved. I've never seen this happen.

I know, can you download this without getting the *.txt extension?

OK I see now. I just need to make sure that every image gets an extension.

HunterTits says

richwicks says

I know, can you download this without getting the *.txt extension?

That one has a .jpg extension in it.

This one does not:

https://patrick.net/post/1346861/2022-07-30-court-rejects-google-s-attempt-to

Under Linux on both Firefox and Brave - that gave me an automatic file extension of *.png and the default file name was: 38ae6f97b04dd387cb9eb02e099d3aa3.png

But it seems like Patrick found the problem. Linux uses a program called "file" which will identify what a file is, maybe it's adding the extension for me. Strangely, file extensions under Linux are by CONVENTION - i.e., they are there for the user's convenience. I can open a file with an extension of .XYZ and it will open correctly 99.9% of the time.

True, it's just a convention and does not seem to actually be used by browsers.

Maybe I don't actually want images to have extensions if it's just for the convenience of downloading and reposting them elsewhere.

In fact, I actively try to block hotlinking of images from patrick.net (including a patrick.net image url on another site) for both resource constraint and legal reasons.

You should use file extensions IMO.

Browsers look at the http response content-type header to determine what type of file it is. Your web server uses the file extension to determine which content-type header to send (unless you use custom configuration). At least in the past, some browsers would use the file extensions to determine the file type in the event that no content-type header was sent, although I recall reading they may have stopped doing it in some scenarios. Sometimes they will try make a last ditch guess depending on usage content: like if the url was referenced from an img element, they will try to use it as an image and so on for other usage contexts, but this can have unexpected results, and many hacks have used this "feature" to get scripts executed if the conditions were right.

If a user downloads and saves a file without an extension, there is no longer any http header or info to tell the next program which opens it what type of file it is. I think most browsers will now add a file extension when saving if it knows the type, but in the past this wasn't the case, and I'm not sure if all browsers always do it now. Having software guess the file type can be dangerous in some scenarios, as its not at all difficult to trick the "guessing algorithm", and has been involved in many exploits over the years. Not to mention the file type can not always be properly guessed, although I don't think this last point is an issue for patnet given that I think it only allows a handful of common media file types, and these tend to have distinct file content signatures that can be guessed well.

Most professional web software I've seen has a small whitelist of allowed file extensions, and only allows the upload of files which use one of those extensions.

Patrick says

Maybe I don't actually want images to have extensions if it's just for the convenience of downloading and reposting them elsewhere.

I don't generally repost, but I save. TRY AND STOP ME!!!

It's interesting you can build up a history by just going to memes. Remember when videos were actually viral and innocent?

https://www.youtube.com/watch?v=nGeKSiCQkPw

original link


There are no naturally occurring viral videos now. It's all artificial.

Hircus says

You should use file extensions IMO.

OK, you convinced me. I'll add extensions.

@HunterTits OK, from now on, images cached from an external url will retain the image suffix, like .jpg

Are you able to match up the @ sign with multiple hits across the border of a space?

If you match with @xxx yyyy send the message to them. If you match with @xxx send to them - in that order.

This seems possible.

Disallow usernames with more than one space, I don't think this would be a problem, but you'd have to verify of course and you'd actively have to prevent it.