Bug or no bug aside, I question whether it is a wise idea to allow such
filenames on the Commons. The point of the Commons is to have a central archive
of files that all Wikimedia projects in all languages can easily use. The vast
majority of Wikimedia users can enter Latin1 characters, whereas many will have
huge problems coping with Hindi or even displaying it properly. I strongly
suggest using an ASCII filename whenever possible if you want your files to be used.

beesley wrote:

The image in question had Hindi text on it, so it's fairly safe to assume that
the people using it would have the ability to use Hindi characters. Those who
can't could simply copy and paste the filename. I don't see this difficulty as a
reason to purposefully disallow Hindi text, which could discourage [[hi:]] users
from using the Commons, preferring instead to retain their Hindi filenames and
upload them directly to the Hindi language projects.

I dont think we want anything else but ASCII in filenames.
Much easier to maintain and spread around.

bugzillas+padREMOVETHISdu wrote:

(In reply to comment #2)

I don't see this difficulty as a
reason to purposefully disallow Hindi text, which could discourage [[hi:]] users
from using the Commons, preferring instead to retain their Hindi filenames and
upload them directly to the Hindi language projects.

I feel that there aren't many images with hindi names around [this is presumably
because Indian languages require complex text layout (CTL) to render properly,
which is a pretty recent concept in the history of computing] and most Hindi
users wouldn't be that discouraged in using Commons [since most of them know
English and would as well rename the image in ASCII, presumably due to the
reason above, computing in Indian languages hasn't caught up much until
recently], at least in the near future. But having said that, what Angela says
may hold true for other languages (with a non-Latin script).

BTW I'm saying this as an Indian who knows Hindi, but of course there might be
hi: Wikipedians who might insist on support for Hindi filenames. Just my 2
[[paise]] :).

Wisdom or no wisdom, if it works at hi it should be working at
commons and I'd like to see the bug fixed. Hemanshu, what
are the file names you've tried to upload? If the file's up at hi,
can you provide a link to it please?

node.ue wrote:

%%%## can be used in URLs for those users whose configurations do not
support Hindi text.

I believe that Unicode-enabled filenames are a must in a project like commons.

(In reply to comment #6)

I believe that Unicode-enabled filenames are a must in a project like commons.

Unicode filenames *are* supported. Without a failing sample file to test with,
however, this particular bug will probably not be located or fixed.

avarab wrote:

I'm marking this bug as invalid as I was able to upload two files to the commons
whos filenames collectively make up the entire Devanagari script (see
http://en.wikipedia.org/wiki/Devanagari#Devanagari_in_Unicode), the characters
in that script are:
ःऄअआइईउऊऋऌऍऎएऐऑऒओऔकखगघङचछजझञटठडढणतथदधनऩपफबभमयरऱलळऴवशषसह़ऽािीुूॉॊोौ्ॐ॒॑क़ख़ग़ज़ड़ढ़फ़य़ॠॡॢॣ।॥०१२३४५६७८९॰

The files I uploaded:
*http://commons.wikimedia.org/wiki/Image:%E0%A4%B5%E0%A4%B6%E0%A4%B7%E0%A4%B8%E0%A4%B9%E0%A4%BC%E0%A4%BD%E0%A4%BE%E0%A4%BF%E0%A5%80%E0%A5%81%E0%A5%82%E0%A5%89%E0%A5%8A%E0%A5%8B%E0%A5%8C%E0%A5%8D%E0%A5%90%E0%A5%92%E0%A5%91%E0%A4%95%E0%A4%BC%E0%A4%96%E0%A4%BC%E0%A4%97%E0%A4%BC%E0%A4%9C%E0%A4%BC%E0%A4%A1%E0%A4%BC%E0%A4%A2%E0%A4%BC%E0%A4%AB%E0%A4%BC%E0%A4%AF%E0%A4%BC%E0%A5%A0%E0%A5%A1%E0%A5%A2%E0%A5%A3%E0%A5%A4%E0%A5%A5%E0%A5%A6%E0%A5%A7%E0%A5%A8%E0%A5%A9%E0%A5%AA%E0%A5%AB%E0%A5%AC%E0%A5%AD%E0%A5%AE%E0%A5%AF%E0%A5%B0.jpeg
*http://commons.wikimedia.org/wiki/Image:%E0%A4%83%E0%A4%84%E0%A4%85%E0%A4%86%E0%A4%87%E0%A4%88%E0%A4%89%E0%A4%8A%E0%A4%8B%E0%A4%8C%E0%A4%8D%E0%A4%8E%E0%A4%8F%E0%A4%90%E0%A4%91%E0%A4%92%E0%A4%93%E0%A4%94%E0%A4%95%E0%A4%96%E0%A4%97%E0%A4%98%E0%A4%99%E0%A4%9A%E0%A4%9B%E0%A4%9C%E0%A4%9D%E0%A4%9E%E0%A4%9F%E0%A4%A0%E0%A4%A1%E0%A4%A2%E0%A4%A3%E0%A4%A4%E0%A4%A5%E0%A4%A6%E0%A4%A7%E0%A4%A8%E0%A4%A9%E0%A4%AA%E0%A4%AB%E0%A4%AC%E0%A4%AD%E0%A4%AE%E0%A4%AF%E0%A4%B0%E0%A4%B1%E0%A4%B2%E0%A4%B3%E0%A4%B4.jpeg

gangleri wrote:

Hallo!

It seems that this bug is related to
Bug 1527: problem on URL with Devanagari characters
This is about the difference of link detection when Unicode precombined
characters / Unicode non-precombined characters are used in links.

please see
http://commons.wikimedia.org/wiki/User:Gangleri/tests/%E0%A4%85%E0%A4%82%E0%A4%97%E0%A5%8D%E0%A4%B0%E0%A5%87%E0%A4%9C%E0%A4%BC%E0%A5%80.png

I used the two variants for uploading the image

http://commons.wikimedia.org/w/index.php?title=Special:Upload&wpDestFile=%E0%A4%85%E0%A4%82%E0%A4%97%E0%A5%8D%E0%A4%B0%E0%A5%87%E0%A4%9C%E0%A4%BC%E0%A5%80.png

http://commons.wikimedia.org/w/index.php?title=Special:Upload&wpDestFile=%E0%A4%85%E0%A4%82%E0%A4%97%E0%A5%8D%E0%A4%B0%E0%A5%87%E0%A5%9B%E0%A5%80.png

1. is using %E0%A4%9C + %E0%A4%BC
2. is using %E0%A5%9B

Ævar's examples are used at the sections "example bugzilla:00830#c8 1" and
"example bugzilla:00830#c8 2".
They display *properly*.

However [[commons:image:अंग्रेज़ी.png]] from the top of the page does *not* display
at all.

To all:

Please reopen the bug if you can confirm this.

To Hemanshu:

Please provide more details about the problem.
Was the problem in "uploading" the files or in "displaying" them?
Please provide links with your answers.
Thanks in advance for the help.

best regards reinhardt [[user:gangleri]]