[GUI Plugin] FanFicFare

[ownedbycats]

Question: If I put the space at the end of a replace_metadata, it gets truncated. How to prevent?

EDIT: Also, the fanfic is inspired by [Restricted Work]. Do I just put always_login to get it to show the title? another edit: Yes

[JimmXinu]

Originally Posted by ownedbycats

Question: If I put the space at the end of a replace_metadata, it gets truncated. How to prevent?

Use \s

Not sure where it's documented other than buried in the defaults.ini replace_metadata comments:

## You can use \s in the replacement to add explicit spaces.  (The config parser
## tends to discard trailing spaces.)

[ownedbycats]

That works, thank you

[tkshorty]

Hi,

when fff finishes download, it clears the marks for all books before applying a mark to the new books (if selected in config).

is there a way to not clear previously marked books when fff finishes?

[JimmXinu]

Originally Posted by tkshorty

Hi,

when fff finishes download, it clears the marks for all books before applying a mark to the new books (if selected in config).

is there a way to not clear previously marked books when fff finishes?

FFF's config, General tab, lower left, there's a checkbox for 'Mark added/updated'.

It's off by default, so you must have turned it on at some point. Possibly to set FFF to report errors--which will clear existing marks, even if there aren't any errors.

I could pretty easily change it to only clear marks when FFF has marks to make.

But would that feel like inconsistent behavior when there are errors and it clears your marks after you've gotten used to it not?

[tkshorty]

Originally Posted by JimmXinu

FFF's config, General tab, lower left, there's a checkbox for 'Mark added/updated'.

It's off by default, so you must have turned it on at some point. Possibly to set FFF to report errors--which will clear existing marks, even if there aren't any errors.

I could pretty easily change it to only clear marks when FFF has marks to make.

But would that feel like inconsistent behavior when there are errors and it clears your marks after you've gotten used to it not?

I don't think a change of the default behaviour is the best way. as you said it may annoy anyone who has gotten used to how it works.

I would however appriciate a raido box or even just a value you can change in personal.ini to not clear previous marks regardless if an error happens or not.

as you lable the marks, I personally don't see a reason to clear previous marks at all, as you can just as well only display the ones marked fff_success and fff_error

[JimmXinu]

Originally Posted by tkshorty

I don't think a change of the default behaviour is the best way. as you said it may annoy anyone who has gotten used to how it works.

I would however appriciate a raido box or even just a value you can change in personal.ini to not clear previous marks regardless if an error happens or not.

as you lable the marks, I personally don't see a reason to clear previous marks at all, as you can just as well only display the ones marked fff_success and fff_error

Ah, so rather than stopping FFF from marking books, you want both FFF marking and your prior marking. (Because otherwise you would just uncheck the FFF marking option.)

Looking at the Calibre code, there is a add_marked_ids() method that was added about 2 years ago(in v6). So Calibre considers it an option now.

Is this something anyone else would find useful?

It's not quite as easy as just calling that instead, since FFF still supports older Calibre versions. Although, there are fewer old version users these days.

My rule of thumb has always been to compare each major version to how many users are on OS 'oth' (ie linux). Dropping v2, 3 & 4 would finally get rid of python2... That may be a project for the new year.

[ownedbycats]

Originally Posted by JimmXinu

Is this something anyone else would find useful?

Yes. A couple of times I went "oh crap, my marked books!" when it was too late to cancel the updates.

[ownedbycats]

Is there a way to minimize the 'download from URLs' dialog? It obscures a decent part of the screen when you're browsing through fic listings looking for interesting stories to grab.

[JimmXinu]

Originally Posted by ownedbycats

Yes. A couple of times I went "oh crap, my marked books!" when it was too late to cancel the updates.

Still seems pretty niche to me, but I'll look into it.

Originally Posted by ownedbycats

Is there a way to minimize the 'download from URLs' dialog? It obscures a decent part of the screen when you're browsing through fic listings looking for interesting stories to grab.

Not that I know of. It's a dialog, not a 'window'.

Uncheck the 'Keep on top' feature and then it won't... stay on top.

[JimmXinu]

Originally Posted by JimmXinu

Ah, so rather than stopping FFF from marking books, you want both FFF marking and your prior marking. (Because otherwise you would just uncheck the FFF marking option.)

Looking at the Calibre code, there is a add_marked_ids() method that was added about 2 years ago(in v6). So Calibre considers it an option now.

I take it back.

The add_marked_ids() method only supports marked:true. It discards both pre-existing and new mark names and calls them all 'true'. Marks used to be only on/off--named Marks where a later add-on.

So the only option I see is to use the db.data.marked_ids dict directly to see what marks already exist and set them again as well.

Looking directly at the data structures instead of using an API is sketchy, even if that's how actions/mark_books.py does it.

Further thought then raises the question: What should happen when a new FFF mark would go on the same book as a pre-existing mark? Which should take priority? The whole point was to preserve pre-existing marks and now we can't guarantee it?

This is the point where I'm thinking it's not worth the effort. Marks are intentionally transient in nature. Use a column value or a Reading List if you want something that is more durable.

[JimmXinu]

New Test Version Posted

2024-12-15
- adapter_fimfictionnet: Implement always_login. Remove unused fail_on_password & do_update_hook settings. #1135

Previously, always tried to login when username/password were set. Currently login isn't working through Flaresolverr, which revealed this.

[vorm]

I observed that there are some problems when downloading images from Xenforo sites:

1 - I at first found this problem in a QQ fic, here is an example:

In this fic ...

Moderator Notice

Mobile Read doesn't allow links to NSFW content. Normally I'd just remove the links, but the logs are all full of NSFW URLs, too.

Maybe report QQ issues in github in future? -- Jim

and

[JimmXinu]

Originally Posted by vorm

I observed that there are some problems when downloading images from Xenforo sites:

I don't think MR sends a notification about edits, so here's a new post to trigger notification for you:

Please don't post NSFW links, mobile read doesn't allow it. I edited your post to remove everything, but I did take a copy to look at your issue. Open a github issue in future for NSFW story issues.

[JimmXinu]

Originally Posted by vorm

I observed that there are some problems when downloading images from Xenforo sites:

I did, however, grab a copy of your issues to look at:

Issue the First

QQ images failing due to lazyload. SB, SV and QQ use the same forum software, but there are differences. Long story short, quite some time ago, SB/SV switched to using a 'data-url' attribute instead of prior 'data-src'. So FFF uses 'data-src', unless there's also a 'data-url', then it uses that.

QQ, otoh, appears to use both, data-src with a link to a proxy on their own server, and data-url to the original source.

In the stories given, those images are either no longer available, or on a server with bad SSL and the QQ proxy cached/masks it.

I can pretty trivially change the code to prioritize data-src over data-url. I don't think that will break anything else. It doesn't in the cases I've checked.

Since I do read SB/SV/QQ, I'm going to run with this change myself a couple days to watch for problems.

Issue the Second
Embeds from Imgur and google drive are not simple images. Too complicated for FFF to handle. Unrelated to prior issue.


Issue the Third
e-shuushuu.net is one of those sites that serves a page instead of an image if it doesn't like the Referer. Can be addressed with personal.ini setting:

force_img_self_referer_regexp:e-shuushuu.net

Unrelated to prior issues.