[damo]

Jdownloader reports plugin defekt for hitomi.la galleries since yesterday.

[Jiaz]

wait for next plugin update

[damo]

Unfortunately, it is broken again.

[Jiaz]

Either cat/mouse or random change. Wait for next plugin update

[Etshy]

Not here to complain, just to say I can confirm issue since today for me (last time I downloaded from hitomi was about one week ago)
As this thread is marked as Solved, should I create a new one ?

[Jiaz]

@Etshy: wait for another update

[Etshy]

Sure

[damo]

Broken again :(

[Cesra250]

Plugin still broken.

[pspzockerscene]

See the prefix of this thread.
You're also free to fix it yourself - we're open source.

-psp-

[Etshy]

Hitomi stopped working since a few days (I didn't try to download for a few days and today every link are "offline" instead of detecting the pages.).

I'll try to see why, but I don't know sh*t about java, so I won't fix it, just try to help find the problem.

EDIT :

Here are my notes from quick look

ProgressController argument used is deprecated (just saying, I don't thing that's the problem)

I guess the "Offline Files" in Hitomi plugin comes from these lines

decryptedLinks.add(createOfflinelink(parameter));
return decryptedLinks;

If that's the case, one possible reason could be these lines

final String js = br.getRegex("src\\s*=\\s*\"([^\"]+" + gallery_id + "\\.js)\"").getMatch(0);
if (js == null) {
     logger.info("Seems like this is no downloadable/supported content");
     decryptedLinks.add(createOfflinelink(parameter));
     return decryptedLinks;
}

Not sure but I guess this code will search for a html tag with the "*gallery_id*.js" as a src attribute ?
From what I see in galleries and reader page this script tag doesn't exist.
On the Reader page, they load "*gallerie_id*.js" from jquery at this url :

**External links are only visible to Support Staff**

the "ltn" part is fixed for prod server

var domain = (/^dev\./.test(document.location.hostname.toString()) ? 'dev' : 'ltn')+'.hitomi.la';

and the rest of the URL too

var result = window.location.href.match(/\/reader\/([0-9]+)\.html/);
var galleryid = result[1];
$.getScript('//'+domain+'/galleries/'+galleryid+'.js' ...

example what I mean with "reader page" : **External links are only visible to Support Staff****External links are only visible to Support Staff**
example what I mean with "galleries page" : **External links are only visible to Support Staff****External links are only visible to Support Staff**

There are also some changes in method "copied" from their js

private String url_from_url_from_hash(String galleryid, Map<String, String> image, String dir, String ext, String base) throws Exception {
    //NEW (JS)
    if ('tn' === base) {
        return url_from_url('**External links are only visible to Support Staff**+dir+'/'+real_full_path_from_hash(image.hash)+'.'+ext, base);
    }
    //ENDNEW
    return url_from_url(url_from_hash(galleryid, image, dir, ext), base);
}

New method "real_full_path_from_hash"

function real_full_path_from_hash(hash) {
        return hash.replace(/^.*(..)(.)$/, '$2/$1/'+hash);
}

private String url_from_hash(String galleryid, Map<String, String> image, String dir, String ext) throws Exception {
	//CHANGE JS - not sure the "split" part is equivalent, guess the java part shoud be just "split(".")" too ?
	ext = ext || dir || image.name.split('.').pop();
	//CHANGE JAVA (existing)
    ext = isNotBlank(ext) ? ext : (isNotBlank(dir) ? dir : last(image.get("name").split("\\.")));
    //ENDCHANGE
    dir = isNotBlank(dir) ? dir : "images";
    return "**External links are only visible to Support Staff** + dir + '/' + full_path_from_hash(image.get("hash")) + '.' + ext;
}

The whole subdomain_from_url function is too different from the JS one in hitomi sources, so I can't say for sure there is no changes

Also it seems the "base" variable is always "a" and "dir" variable is either "webp" or "avif" now, but seems "webp" will always work.
(from what I can see in JDownloader sources, it was null before)
(see in reader.js, search for "url_from_url_from_hash(", the only calls without "a" as base variable



That was a really quick look though, I hope I wasn't too confusing and this will help.

[Jiaz]

@Etshy:They have changed again, wait for next plugin update. Can't promise jpeg support as they move to webp and avif image format.
Please check again with next update. I'm sorry but they stopped jpg and only have avif and webp. in case browser doesn't support webp, they use javascript webp decoder

[Etshy]

Yeah we already saw that they stopped jpeg support a few months ago and was transitionning to webp.
from their JS I guess they completely change from jpeg to webp as almost everything seems to use webp.

Thanks a lot for the quick update

[Jiaz]

Originally Posted by Etshy

Thanks a lot for the quick update

Wait for update and thank me once it's working for you I've updated plugin to prefer webp over avif, okay?

[Etshy]

Okay I'll wait to really thank then ^^
For me webp is fine, no problem.

[Jiaz]

@Etshy: Update will be available in about 5 mins

[Etshy]

Seems they changed something else about the url generation
All files appears as not Found and have 403 when i go to them
https://i.imgur.com/JeBkhY5.png
Example url generated by jdownloader : **External links are only visible to Support Staff****External links are only visible to Support Staff**

Even on their reader page ( **External links are only visible to Support Staff****External links are only visible to Support Staff** ) the webp seems to be in 403, so that's a bit weird (their json for this manga show this manga have webp though.)
**External links are only visible to Support Staff****External links are only visible to Support Staff**

The subdomain seems to be wrongly guessed too

I'll make some more research tonight.

Edit :

webp url from Jdownloader (not working) : **External links are only visible to Support Staff****External links are only visible to Support Staff**
webp url from hitomi donwloader : **External links are only visible to Support Staff****External links are only visible to Support Staff**


I found something about the "gg" variable, a part of the process that find the subdomain and first path part.
there is a file "gg.js" that doesn't seem to be loaded in Jdownloader.
It's just an object with with 3 properties/methods (it should be simple to remake them I think)

Some process about finding a number and then transforming this number into a letter to generate the subdomain

That should fix the problem with the subdomain

And then there is another problem with the first part of the path (after the '/webp')

in js

function full_path_from_hash(hash) {
        return gg.b+gg.s(hash)+'/'+hash;
}

//gg methods
s: function(h) { var m = /(..)(.)$/.exec(h); return parseInt(m[2]+m[1], 16).toString(10); },
b: '1649948402/'

both errors seems to be related to that "gg.js".
I guess this is not loaded, or it's loading some old script from their server
The actual gg.js is here : **External links are only visible to Support Staff****External links are only visible to Support Staff**

I guess the "engine" in HitomiPlugin is executing the js from the page it loaded ?
so just loading the correct gg.js should be enough I think.

[Jiaz]

@Etshy: I will have to check again. My example links did work fine. I'm sure something I just have overseen

[Jiaz]

Originally Posted by Etshy

I guess the "engine" in HitomiPlugin is executing the js from the page it loaded ?
so just loading the correct gg.js should be enough I think.

Plugin is already using the same one

[Etshy]

I really don't know what's wrong with the URL generation whthin JDownloader.
But if you can exec the js script, why not call their "url_from_url_from_hash" directly ?

if (haswebp == 1) {
                    url = url_from_url_from_hash(gallery_id, picInfo, "webp", null, "a");
                } else if (hasavif == 1) {
                    url = url_from_url_from_hash(gallery_id, picInfo, "avif", null, "a");
                } else {
                    url = url_from_url_from_hash(gallery_id, picInfo, null, null, null);
                }

Here you could just exec their function, no ?

Not sure if the following is good I wrote that from example in the source

//create a variable containing the object image in js
engine.eval("var image=JSON.parse(" + JavaObjectToJSON(picInfo)+ ");");
if (haswebp == 1) {
    engine.eval("var url = url_from_url_from_hash(" + gallery_id + ",image, 'webp', null, 'a');");
} else if (hasavif == 1) {
    engine.eval("var url = url_from_url_from_hash(" + gallery_id + ",image, 'avif', null, 'a');");
} else {
    engine.eval("var url = url_from_url_from_hash(" + gallery_id + ",image);");
}

final String result = engine.get("url").toString();

Honestly I have no idea if it would work. But if the js is loaded in the "engine" it should work I guess.