Do not escape audio src path

[ltguillaume]

Do not escape audio src path

[GM-Script-Writer-62850]

Without any escaping as long as there is not a # it works
Using encodeURIComponet break everything, but if you were to unencode / it would probably work
Using encodeURI break nothing, works just as well as not escaping
Using escape breaks spacial Unicode characters

encodeURI does not touch these:
A-Z a-z 0-9 ; , / ? : @ & = + $ - _ . ! ~ * ' ( ) #
the only characters that look to break this are # and ?

[ltguillaume]

Does it do any harm then? I have a folder # for all artist names starting with numbers or special characters. If I don't use encodeURIComponent, it won't play (it sees it as an anchor, instead of a folder). Of course, I can only replace #, but I wasn't sure if any other special chars did the same.

[GM-Script-Writer-62850]

The problem it is escapes forward slashes so it will not work with sub folders
this is form my firefox console

[ltguillaume]

Forward slashes are used on a Windows HTTP server?

[GM-Script-Writer-62850]

I am using a apache2 on my linux server: https://imgur.com/a/5CtPA

Did you have any issues with my version working?

[GM-Script-Writer-62850]

does using the escape function instead work for you, that did not break it for me?
audio.src = escape(path);

[ltguillaume]

Strange, I thought the behavior of the deprecated escape() would mimic encodeURI() or encodeURIComponent(), but apparently it's a different flavor altogether. escape() works just fine over here, so if it also works in your use case, it's the best option. Thanks!

Apart from this, I also need to add a few lines of error handling when playback fails. I just want it to mark the non-played item and continue with the next, since a party player that stops playing instantly becomes a party pooper.

[GM-Script-Writer-62850]

audio.src = encodeURI(path); also works for me

[ltguillaume]

Not for me in the # case, so escape() is best :-)

[GM-Script-Writer-62850]

Did you have this issue is my version?

[ltguillaume]

Yeah, I did. Back then, I just added s.src=file.replace('#', '%23');.

[ltguillaume]

b461afc

[GM-Script-Writer-62850]

• Prevent encoding of \ in audio.src for Apache servers

that is the wrong slash

[ltguillaume]

Lolz, that's odd. I could have sworn I saw actual backslashes when I checked your post. I didn't understand, that's why I asked whether you were running a server on Windows, the only theoretical case I could fathom where backslashes could be used! Anyway, all the more reason I shouldn't code on a 10" screen, it issues problems that aren't there :-P Thanks again, escape() seems to be working fine in a few test cases I did.

[GM-Script-Writer-62850]

escape seems to break it with some unicode characters, here is one for you to try: 月

[GM-Script-Writer-62850]

Using encodeURI seems to to not break anything here
encodeURIComponent is used for parameters like this
http://example.com/path/to/file?opt='+encodeURIComponent(string)+'&'+encodeURIComponent(string2)

[GM-Script-Writer-62850]

I guess we should use this:
encodeURI(FILE_PATH).replace(/#/g,'%23');

[ltguillaume]

And it does work without any escaping? Then I'll just create a function that replaces `#` with `%23` and go from there (if there are any other characters that _do_ need escaping. Thanks so much!

[ltguillaume]

Wow, that's a thorough investigation. (Probably) definitively fixed in 1.0.2 now :-)

[GM-Script-Writer-62850]

s.replace('#', '%23').replace('?', '%3F');
but what if there is more than one?

function safeURI(src){
   return encodeURI(src).replace(/#/g,'%23').replace(/\?/g,'%3F');
}

I feel like there is a way to use replace only once...

[ltguillaume]

Dang, I need to get some sleep, not focused at all anymore.,, Either way, I feel using RegExp will inadvertedly introduce a whole new can of worms with RegExp's metacharacters, so simply using return s.split('#').join('%23').split('?').join('%3F'); is best for this occasion imho. I found my suspicion is documented on StackOverflow, too.

This solution isn't fast, but occurrences will be very limited anyway, so it shouldn't delay the script much at all.

[GM-Script-Writer-62850]

StackOverflow link please

[ltguillaume]

https://stackoverflow.com/questions/1144783/how-to-replace-all-occurrences-of-a-string-in-javascript#comment27121391_1144788

There's a way around it, of course, but I'd rather keep the code more legible. I genuinely hate RegExp, because I keep being awed by its power while stunned and frustrated by its pickyness and lack of intuitiveness :-P

[GM-Script-Writer-62850]

Knew there was a way to only use replace once

function safeURI(src){
   return encodeURI(src).replace(/[(#\?]/g,function(char){return escape(char);});
}

There are 10 kinds of people who understand regular expressions: them and us

[ltguillaume]

Cool! And using encodeURI() just to be on the safe side? Just tried your updated project, looks awesome! Tabbed interface is nice. It's working great in Vivaldi (Chrome), but in Waterfox I get the error "config is null" with a reference to the line where you check if config is undefined. Script stops and nothing works.

[GM-Script-Writer-62850]

encodeURI is to be on the safe side
encodeURI is intended to encode a entire url, it does not target ?, =, &, and # as these special url chars
? is used to send parameters to a server
= is used to separate parameters names from values
& is used to separate separate parameters
# is used for scrolling to elements on a page based on element IDs
I think we could escape both = and &, but browsers know if there has not been a ? symbol they are not being used for parameters
function safeURI(src){return encodeURI(src).replace(/[(\?=&#]/g,function(char){return escape(char);});}
In theory we should be able to use encode uri component and unescape every / if we do not want to escape a few more chars our self
function safeURI(src){return encodeURIComponent(src).replace(/%2F/g,'/');}
If this is used on a full link you need to decode : also
function safeURI(src){return encodeURIComponent(src).replace(/(%2F|%3A)/g,function(str){return decodeURIComponent(str);}); }

---

function safeURI(src){
   // escape additional chars method
   return encodeURI(src).replace(/[(\?=&#]/g,function(char){return escape(char);});
   // unescape necessary character
   return encodeURIComponent(src).replace(/%2F/g,'/');
   // unescape multiple chars to be safe
   return encodeURIComponent(src).replace(/(%2F|%3A)/g,function(str){return decodeURIComponent(str);});
}

And with that my regex cheat sheet has been consumed

[GM-Script-Writer-62850]

config is null; easy fix; only need to add 4 chars
I do not think that was a waterfox issue, it was a there is no config data yet issue

[ltguillaume]

True, but there wasn't any in Vivaldi either.

[GM-Script-Writer-62850]

firefox should behave the same, the only way for it not to is if the config option has not been set

it a very inconvenient to debug that part as it is set to save the config on page leave, so even if you clear it then reload it saves it and you do not get a cleran config

if you used the same config name i did in your version on the same server both pages share the same config data, if you used Vivaldi on your project then loaded mine on the same server you had a exist config that was not null

also according to a scratch pad test firefox would get this error also as this returns null:
JSON.parse(localStorage.getItem("someFakeStorageName-"+Math.random()))
if that does not return null the browser does not follow the standard

[ltguillaume]

I waited for Cookies Exterminator to clean up cookies & LS, but I guess it didn't take those instances. Sorry to have misinformed you.

Oh and no, I'm not using the same config name in my project.