P: Win LR doesn't correctly open catalogs exported by Mac LR if the folders contain non-ASCII chars

[johnrellis]

Windows LR doesn't correctly open a catalog folder exported with Mac LR with File > Export As Catalog, Export Negative Files, if the Mac catalog folder contains subfolders with non-ASCII characters (e.g. á or é).

To reproduce on LR 14.2 / Mac OS 15.2, Windows 11:

1. Download this .zip file containing a catalog and photos to your Mac desktop:

https://www.dropbox.com/scl/fi/stbx8iyfekqk9xkuri3fi/accents.2025-02-14.zip?rlkey=xt283tc32rjarroeu4...

It contains a catalog created with Mac LR by doing FIle > Export As Catalog, Export Negative Files.

2. Unzip and open that catalog in Mac LR. Observe that it contains four photos in two folders and none of the folders or photos are marked missing:

2. Download the .zip file to your Windows desktop.

3. Unzip the .zip using these exact commands:

cd Desktop
tar xfzv accents.2025-02-14.zip

Do not use the Windows ZIP built-in to File Explorer. It still doesn't correctly handle paths containing UTF-8 characters (hey, it's only 2025).

4. In Windows File Explorer, verify that the extracted catalog folder "accents" has this structure, the same as in the original catalog folder on Mac:

accents
    accents.lrcat
    pics
        a
            e.jpg
            é.jpg
        á
            e.jpg
            é.jpg

5. Open the "accents" catalog.

6. Click on All Photographs and observe that all four photos and the three folders are marked as missing (incorrect):

7. In the Folders panel, click on "pics" and observe that "pics" and "a" are no longer marked as missing, while three of the photos are missing and one is not (all incorrect):

8. Click on All Photographs, and observe that "pics" and "a" are marked as missing again. You may have to adjust the thumbnail size larger or smaller a few times until that happens. The erratic behavior here clearly shows LR is confused.

9. Right-click on "pics" and do Find Missing Folder. Navigate to Desktop\accents\pics. Observe that "pics" and "a" are no longer marked missing (correct) but "á" is still marked missing (incorrect). Observe that the photos in "a" are no longer missing (correct):

10. Right-click on "á" and do Find Missing Folder. Navigate to Desktop\accents\pics\á. Observe that "á" and its two photos are no longer marked missing (correct):

[seb.lebreton]

I saw this first a few months ago with LR13 and it is still reproducible with LR 14.1.1.

Steps to reproduce:

1) create a catalog on mac, and place it at the same level as a folder for picture files. In this folder, create subfolders with or without accentuated characters (e.g. 'é', 'à', 'è') and add some picture files (with or withour accentuated characters). Import the main pictures folder in the LR catalog.

2) copy the catalog, associated LR files and the main pictures folder on a usb drive and open it on windows, with the same LR version

Expected result:

All the files and folders are successfully found since they are at the same level as the catalog file

Actual result:

Only the files and folders without accentuated characters are found.

The folders with accentuated characters have a "?" and corresponding images have a "!" icon

With a right click > "Find missing folder..." I can simply select the folder on my usb drive and it will "reconnect" it without having to rename it, but I have to do it manually for each folder or file missing.

Notes:

- Quitting the catalog, reopening it on my Mac, making changes (so that the catalog is saved again), and going back to windows doesn't break the connections I have previously restored.

- Creating a folder on windows with similar characters opens up fine on Mac

- I mentioned 'é', 'à', 'è' as the main characters I used, I haven't tried everything else but I suspect ç ù and maybe some special characters such as œ could trigger the same bug ?

For me this bug is problematic, as I'm working on a mac computer at home, but I have a windows laptop for my trips, so I have to fix it regularly for all the new folders I've created since the last trip.

[seb.lebreton]

I think I posted this in Bugs, but I see it is in Discussion, so maybe I messed up, or somebody moved it here ?

Either way, I'm pretty sure this is a bug, so can it be moved it (back) to Bugs ?

[richardplondon]

There is a moderation process needed I believe, before an issue will appear formally as a Bug - including checks on reproducibility, and on whether enough technical info has been given. So you may get some follow-up questions first.

[richardplondon]

Maybe try avoiding this character (#) and see if that makes a difference? It does sit within that grey area between universally OK, and problematic. I would not say the same thing about the accented characters themselves: these are normal and well tested AFAIK.

[Uninformed speculation]: if that (#) character had raised some error while interpreting these image paths, perhaps LrC could be reverting to just a basic character set for safety - and that tighter interpretation might reject any expanded Unicode characters found.

[Per Berntsen]

A moderator has moved your post to Discussions because this is most likely not a bug.

I'm on Windows, and created a folder named æ'øåéç and imported an image from it. LrC has no problems with this folder.

2) copy the catalog, associated LR files and the main pictures folder on a usb drive

If this is a thumb drive, it might be the cause of the issue. Thumb drives are known to be unreliable.

In the past, the general advice was to only use characters a-z, numbers 0-9, hyphens and underscores in file and folder names, especially on the Internet.

Other characters were considered "illegal", and Internet Explorer would not display images with illegal characters in the file name.

This has obviously changed, but I think I saw a post recently about LrC complaining about illegal characters in the folder path on a Mac. 

Maybe @johnrellis can tell us something about the status of illegal characters.

[seb.lebreton]

Hi all

@richardplondon thanks for the details on your speculation, as it helped me understand what you had in mind. I think it's unlikely though, as on other catalogs I also had this issue in my main folder (e.g. 2023; 2024, not 2023 #WIP# or 2024 #WIP#) which means it doesn't matter if there is a "#" in the path. Maybe this character can still mess something up when scanning the rest of the file tree at launch, but I'm very doubtful. Even then, if this ends up being part of the root cause for my behavior, I still feel like this would be a bug too, that is worth investigating.

@Per Berntsen  it only happens when the catalog was created on mac, and moved to windows. As I mentioned in my post, I have tried creating such folder on the pc, and it's fine, even when moving the catalog back to the mac, and back again to the pc. For me, it only happens the first time LrC tries to read such characters in a folder or filename that was created on mac.

I'm pretty sure it has nothing to do with my usb (SSD) drive, because in my latest try with 14.1.1 I did load the catalog from usb, but in previous instances I have either copied it from USB to my windows internal drive and loaded from there, and also copied it over my network from the mac to the pc (probably did it with usb Hard Drives too before, by the way)

Regarding the use of only the basic characters, as you said, this shouldn't be an issue now. If it is, then I feel like it should be considered as a bug and logged accordingly, even though it may not get the highest priority for the product owner. I have a found a thread about illegal characters in keywords, and another one about "/" in filenames, but that's it

Hope this can be reproduced to confirm status and then hope for the best in therms of prioritization.

Thanks

[johnrellis]

@Rikk Flohr: Photography, there's a related acknowledged bug for going the other direction, exporting from Windows to Mac:

https://community.adobe.com/t5/lightroom-classic-bugs/p-mac-catalogs-exported-by-win-when-source-fol...

in case you want to merge this bug report into that and change the subject line.

[johnrellis]

I filed a bug report about this, exporting catalogs from Mac to Windows when subfolders contain non-ASCII characters:

https://community.adobe.com/t5/lightroom-classic-bugs/windows-lr-doesn-t-correctly-open-catalogs-exp...

Adobe previously acknowledged a similar bug exporting catalogs from Windows to Mac:

https://community.adobe.com/t5/lightroom-classic-bugs/p-mac-catalogs-exported-by-win-when-source-fol...

[johnrellis]

Moderators, @Rikk Flohr: Photography, please merge with the bug report:

https://community.adobe.com/t5/lightroom-classic-bugs/windows-lr-doesn-t-correctly-open-catalogs-exp...

[Rikk Flohr: Photography]

Thanks @johnrellis 
I've cross-referenced with the previous bug and alerted the team.