Piwigo.org

exogadget · 2020-04-23 16:43:23

Hello/Hi/Greetings,

so friends, after few hours surfing the forum and google I see that Piwigo doesn't work with non-English letters in the file/folder names, am I right? I would much appreciate you if you correct my belief.

What is my problem exactly?
I have a lot of folders (and some files) on my NAS Synology (tested not only on Synology but on usual Windows folder) with Cyrillic letters in the names. I vote for English naming, but not only me use these folders. Rename all of them is not proper solution, better I'll find another soft.

These folders are symlinked (mount --bind) to "gallery" folder of Piwigo. When I try to sync them I've got the message:
"[./galleries/photoalbum/день_рожденья] PWG-UPDATE-1 (wrong filename)"

I tried to add this param to config.inc.php in /local/config/ but without success, the problem wasn't solved:
$conf['sync_chars_regex'] = '/^[[:alpha:][:digit:]-_.\sйцукенгшщзхїфівапролджєячсмитьбюЙЦУКЕНГШЩЗХЇФІВАПРОЛДЖЄЯЧСМИТЬБЮёЁъЪыЫэЭ!\&\/\°\'\,]+$/';

If I create Cyrillic album in Piwigo, of course, it works properly, as physically it isn't Cyrillic folder in structure, just a name in DB.

Do you know any solution?

Piwigo version: 2.10.2
PHP version: 7.3.11
MySQL version: 5.5.62-MariaDB
Piwigo URL: http://sorry it's closed for all

Zentalquabula · 2020-04-23 17:47:33

Operating system? What locale settings do you have? I assume Piwigo like all modern software requires UTF-8 in the file system.

exogadget · 2020-04-24 11:27:53

Zentalquabula wrote:
Operating system? What locale settings do you have? I assume Piwigo like all modern software requires UTF-8 in the file system.

Yes, there is UTF-8, all folders were created in Windows 10 and local is Ukrainian which is very similar to Russian.

erAck · 2020-04-24 13:20:19

As $conf['sync_chars_regex'] is used with the PHP preg_match() function you may have to append the u modifier to make that work. Make sure the editor you use writes file content in UTF-8 encoding and that the web server process also runs in UTF-8 locale context. It might also unlikely be your PHP wasn't compiled with UTF-8 for PCRE, in that case prepend a (*UTF8) to the regex. You can also try to express the Cyrillic characters as Unicode code points, e.g. й would be \x{0439} and so on. Instead of listing each individually you can also use a range, e.g. \x{0439}-\x{044E} or whatever suits there.
See https://www.php.net/manual/en/function.preg-match.php

exogadget · 2020-04-27 14:06:31

These modifications didn't solve the issue.

setlocale(LC_ALL, 'ru_RU.UTF-8');
setlocale(LC_NUMERIC, 'C');
$conf['sync_chars_regex'] = '/^[[:alpha:][:digit:]-_.\s\x{0410}-\x{044F}\&\/\°\'\,]+$/';

I'm sorry, I'm not good in PHP, what did I do wrong?

erAck · 2020-04-28 00:30:32

The actual Cyrillic script Unicode range is larger than just \x{0410}-\x{044F} so probably you're missing characters there. For example these characters of your previous expression are not in that range:
ЁЄІЇёєії
which are 0x401 0x404 0x406 0x407 0x451 0x454 0x456 0x457
There are several Cyrillic script blocks (i.e. also Supplement, Extended-A,-B,-C), you seem to use range U+0400 to U+045F of the block U+0400 to U+04FF, see https://r12a.github.io/uniview/?block=cyrillic

You can test your expression against single file names at https://regex101.com/ with the PCRE (PHP) flavor.

Btw, the \ backslash escape for the single characters \&\/\°\'\, is not needed in the character list, except for the / forward slash, it just makes the expression less readable. That could be ()&°',\/ instead.

exogadget · 2020-06-10 12:12:39

erAck wrote:
The actual Cyrillic script Unicode range is larger than just \x{0410}-\x{044F} so probably you're missing characters there. For example these characters of your previous expression are not in that
...
Btw, the \ backslash escape for the single characters \&\/\°\'\, is not needed in the character list, except for the / forward slash, it just makes the expression less readable. That could be ()&°',\/ instead.

Done! Thanks much, erAck!

But I did some additional steps:
1) set collation to utf8mb4_unicode_ci for DB;
2) config.inc.php was changed with:

setlocale(LC_ALL, 'ru_RU.UTF-8');
setlocale(LC_NUMERIC, 'C');
$conf['sync_chars_regex'] = '/^[[:alpha:][:digit:]\_.\s\x{0400}-\x{04FF}()&\/°\',!\-]+$/u';

'u' at the end, \- and \' were the stumbling stones.

Now everything works fine.

erAck · 2020-06-10 18:02:54

Btw, you should remove the \' apostrophe, at least with directory names it does not work and may or may not (or only partially) work with file names, see [Forum, topic 30356] Directory Names with ' in.

exogadget · 2020-06-10 20:01:38

This sign is used in Ukrainian, e.g. Кам'яна, say frankly that isn't a real apostrophe, but thank for your help!
This char could be used in FS file naming.

erAck · 2020-06-10 23:01:36

Nevertheless the character you used is
U+0027 ' APOSTROPHE
however the context might be in your language. The apostrophe does confuse some processing hence should not appear in the regex.

exogadget · 2020-06-12 08:24:15

You're right. It causes. Checked on filenames with \'. Excluded it from naming.

Viva · 2022-12-17 17:31:17

Guys, I've faced with the same problem. I've made all the recommendations, but sync works strange.
For example
I have album "2002/Дорога домой" (2002 is a top level album)
Synchronisation add this album with no errors, but when I open it in album it called "2002/ домой"
It removes first Cyrillic word. And if I try to sync it next time it removes this album and then add it one more time. And the situation repeats.

If I put any digit or symbol before Cyrillic word - everything is like a charm.

What should I do?

p.s. Piwigo 13.4.0

Last edited by Viva (2022-12-17 17:31:56)

a33mail2013 · 2023-02-15 12:41:01

exogadget wrote:
Done! Thanks much, erAck!

But I did some additional steps:
1) set collation to utf8mb4_unicode_ci for DB;
2) config.inc.php was changed with:

setlocale(LC_ALL, 'ru_RU.UTF-8');
setlocale(LC_NUMERIC, 'C');
$conf['sync_chars_regex'] = '/^[[:alpha:][:digit:]\_.\s\x{0400}-\x{04FF}()&\/°\',!\-]+$/u';

'u' at the end, \- and \' were the stumbling stones.

Now everything works fine.

[s]I've made the same steps, DB itself and all tables are utf8mb4_unicode_ci now.
But Piwigo can't even add a standalone pic outside cyrillic named folder.
[/s]

My fault. Apologies for my post.
Everything works fine now and no need to change collation to utf8mb4_unicode_ci for DB
Piwigo 13.5.0
Installed on 14 February 2023, 14 hours ago
Operating system: WINNT - Windows 7 x64
PHP: 7.4.33
MySQL: 5.5.27
Graphics Library: GD bundled (2.1.0 compatible)

Last edited by a33mail2013 (2023-02-16 22:25:43)

Piwigo.org

Announcement

#1 2020-04-23 16:43:23

Cyrillic and other unconventional letters in name of symlinked gallery

#2 2020-04-23 17:47:33

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#3 2020-04-24 11:27:53

Re: Cyrillic and other unconventional letters in name of symlinked gallery

Zentalquabula wrote:

#4 2020-04-24 13:20:19

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#5 2020-04-27 14:06:31

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#6 2020-04-28 00:30:32

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#7 2020-06-10 12:12:39

Re: Cyrillic and other unconventional letters in name of symlinked gallery

erAck wrote:

#8 2020-06-10 18:02:54

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#9 2020-06-10 20:01:38

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#10 2020-06-10 23:01:36

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#11 2020-06-12 08:24:15

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#12 2022-12-17 17:31:17

Re: Cyrillic and other unconventional letters in name of symlinked gallery

#13 2023-02-15 12:41:01

Re: Cyrillic and other unconventional letters in name of symlinked gallery

exogadget wrote:

Board footer