ascii file-renaming

[suppy]

in relation to this: http://www.anidb.net/forum/viewtopic.php?t=4681 I would like to request an option to limit the character-set used in AOM's automatic file-renaming to strict ASCII, because some of the weird characters do get through this bug, and they can cause problems when downloading AOM-renamed files through FTP yes, I'm having this problem right now (getting Princess Tutu ... ep 2 (Akt 02 - Kapitel des Eies: Pieces of the Heart ~ Schwanensee: Scène finale) wouldn't download, and I believe ep 7 (Akt 07 - Kapitel des Eies: Raven Princess ~ An der schönen blauen Donau), 30 (Akt 22 - Kapitel des Junges: Crown of Stone ~ Das große Tor von Kiew (Part 1)), 31 (Akt 22 - Kapitel des Junges: Crown of Stone ~ Das große Tor von Kiew (Part 2)) and 38 (Akt 26 - Kapitel des Junges: Finale ~ Der Nußknacker) will be hard too ... atleast my ftp client seems to indicate that they will)

luckily, the guy can change the filenames, but AOM will rename them again...

this problem does seem to be a bit inconsistent though, cause it had no problem with either episode 1 (Akt 01 - Kapitel des Eies: Ahiru and the Prince ~ Der Nußknacker: Blumenwalzer) or episode 3 (Akt 03 - Kapitel des Eies: A Princess`s Oath ~ Dornröschen: Panorama) ... though the episode-name in the filename was cut off right before the weird characters in them.

I would say that the difference is that in episode 1 (the one that is working correctly) the episode name had non-ascii characters entered as a html-escaped sequence, while in ep 2 (the one that didn't work), the non-ascii character was entered in (I suppose) ISO 8859-1. The standard way is to enter it as html-escaped sequences IIRC, and I would creq it, but there is already another creq pending (for another matter from the looks of it), so I can't.

I'm not sure whether this should really be a feature request or a bug-report, but having the option of limiting the character-set to only ascii, should be considered a feature I guess.

[PetriW]

AOM will never support this for the simple reason that I do not support the english only thinking ascii promotes. If this is a problem for some of your software then I suggest you ask them to fix it.
Entering html escaped text will only make aom convert them into the correct sign in future versions.

While I kind of understand this might be an issue for some people windows has supported unicode for over 10 years, it's about time the ones who don't support it fix their stuff. This is a problem that should lie only on those who do not support unicode.

[epoximator]

but this could be a possible solution

[Rar]

It'd be easy to provide a mapping of titles to ascii as an option. Could take a unicode text file and perform a find/replace on the contents, and turn anything else into * or something to get ascii only filenames.

Code: Select all

éèê
e

ß
ss

ö
oe 

: 
 - 

etc... For that matter, you could allow any kind of custom find/replace, and provide a few different mappings for different ones (a windows-os one would be useful, for instance).

Rar

[egg]

You can use WebAOM, then you can define your mappings, here are mine:

Code: Select all

true	`	'
true	"	'
true	<	
true	>	
true	|	
true	/	 - 
true	:	 - 
true	\	 - 
true	?	
true	*	
true	é	e
true	é	e
true	’	'
true	  	 
true	  	 
true	∞	 

The ones that look blank are multiple spaces to a single space.

[suppy]

While I kind of understand this might be an issue for some people windows has supported unicode for over 10 years, it's about time the ones who don't support it fix their stuff. This is a problem that should lie only on those who do not support unicode.

yes. blame it on poor programming of the programs being used. the FTP protocol is based on ASCII, so it'd be kindof ... weird ... to make an ftp client and/or server that supported unicode. I'll suggest a total rewrite of the FTP protocol while I'm at it, shall I?

Edit: refer to http://www.faqs.org/rfcs/rfc959.html section 5.3.1 and 5.3.2 for confirmation that filenames can only be specified by

any of the 128 ASCII characters except <CR> and <LF>

[PetriW]

While I kind of understand this might be an issue for some people windows has supported unicode for over 10 years, it's about time the ones who don't support it fix their stuff. This is a problem that should lie only on those who do not support unicode.

yes. blame it on poor programming of the programs being used. the FTP protocol is based on ASCII, so it'd be kindof ... weird ... to make an ftp client and/or server that supported unicode. I'll suggest a total rewrite of the FTP protocol while I'm at it, shall I?

Edit: refer to http://www.faqs.org/rfcs/rfc959.html section 5.3.1 and 5.3.2 for confirmation that filenames can only be specified by

any of the 128 ASCII characters except <CR> and <LF>

Hi, if the server was a decent one it would convert the filenames to something understandable by the ftp protocol. Your problem is most likely caused by it no following the very thing you quoted above.
This is still a limitation of your ftp client / server. Not a limitation in aom / windows / linux.

[Rar]

But... there's not a reason not to add a find/replace patterns feature request for AOM on the tracker, right?

Rar

[PetriW]

But... there's not a reason not to add a find/replace patterns feature request for AOM on the tracker, right?

There's already a request present on the tracker to be able to replace specific characters with other characters like in WebAOM, this will be done.
There will be no generic ascii convert however.