Postal code (zip_code) can't hold codes starting with zero

[marcopenhacking]

Tested with 4.2.99:

• Filled in a zip code 05555
• This was transformed to 5555
• In Germany there are codes starting with zero

contact

street:
zip_code: 5555
city:
country:
email: test@fc.test
phone: 

[wvengen]

Thanks for reporting! This requires some changes to the configuration system, I'm afraid.

As a workaround, you could a space in front of the zero. I documented this on the wiki as well.

[RayOei]

I can't seem to reproduce this issue. The app_config.yml uses a string for the zip.
Also changing the zip in configuration in admin/config seems to work just fine.

[lentschi]

Also changing the zip in configuration in admin/config seems to work just fine.

@RayOei I can still reproduce the issue in the current upstream main: When entering a number with a leading zero, that digit simply disappears on save. (Have you tried this in the upstream version or in a fork? Which db flavor were you using?)

[RayOei]

I was trying this on the article PR. The 0 is still there in my zip code.
It is a MariaDB instance (v10.11.8 on Ubuntu 24.04.1)

[lentschi]

I was trying this on the article PR.

You mean Article versioning, extended units and alternative article sharing (#1058) #1073?

I've upgraded my MariaDB to v10.11.8 (from the currently configured foodsoft standard 10.5.x) and it's still happening for me in the same branch.

I looked into the code and wvengen is right: To be able to set a postal code with a leading 0 one has to put at least one non-digit-character in to the postal_code input.

Else this regex condition will transform it into an integer, which obviously cannot store leading zeros.

This should be independent of database & locale, so are you sure, you hadn't just accidentally added a space somewhere in your test, @RayOei ?

A kind of ugly quick fix could be changing this regex to something like ^[-+1-9][-+0-9]*$ (and the float below similarly to ^[-+1-9.][-+0-9.]*([eE][-+0-9]+)?$). However, a much cleaner solution would be to re-think the config system as wvengen indicated (more effort though.)