Albums: Replace ASCII double quotes in titles with Unicode quotes instead of single quotes

[dror3go]

Following #1196 where we enabled keywords with single/double quotation marks, PhotoPrism currently converts double quotation marks in album titles with single quotation marks.
Though not a high priority bug, it feels weird for users in languages which make use of double quotation marks for acronyms.

[lastzero]

Think that was a precaution measure in the aftermath of the log4j debacle. Double quotes would also simplify code injection and may not be supported on many filesystems in case you want to use album names for zip files or as folder names. As such, it's a decision we made (for now) and not a bug. Feel free to help with improving our docs so that it does not look as that happened by mistake and should be reported as a bug.

[dror3go]

While I understand the technical rationale - as an end user I still find it to be a bug.
Though not a German speaker myself, I think it's a little bit like if German letters such as Ä and Ö would've been replaced with A and O.
As I mentioned in the original post - I think it's not a high priority issue, but still an issue nonetheless. I feel that labeling it as documentation issue is saying it's WONTFIX and I should expect this behavior to remain as-is.

Adding a note on the docs wouldn't improve anything from my perspective - I just don't think that an end user should be expected to read the docs just like I'm not reading manual guides of a new TV for example. Ideally things should work "as expected".

[lastzero]

Would have been replaced? Where? How?

[dror3go]

They're not, that's my point.
As an end user - would you think it was a bug if PhotoPrism replaced the first two letters to "OA"?

[lastzero]

בּד״א
works as well
״ is not the same as "

[lastzero]

All of the examples I found on https://www.myjewishlearning.com/article/a-guide-to-jewish-acronyms-and-abbreviations/ can be used as album names, just not the ASCII " which would require escaping and a lot of security checks and which technically is not the right character for RTL languages either as it is LTR.

[lastzero]

Would you be happy if we convert " to a proper Unicode character?

[lastzero]

I have started a development preview build for testing. It should be ready within an hour. You can then also use our demo to test it.

[lastzero]

See also:

• https://en.wikipedia.org/wiki/Gershayim
• https://www.fileformat.info/info/unicode/char/05f4/index.htm

[dror3go]

בּד״א works as well ״ is not the same as "

Yes, I'm well aware :)

Thing is that de-facto nobody other than Apple devices is using this character, since it's not exposed to users on physical and non-physical keyboards.
Even worse - Apple automatically converts " to ״ even when it's clearly wrong, like in code-related things. Imagine you send your teammate a snippet of a log which includes strings with quotes - your teammate will receive a broken text which can't be copy-pasted to dev-related tools. The best example would be a JSON object.

I personally prefers these quotes over a single quote, but would like to stress that a search for an album with straight quotes must include in the results albums with curly quotes in their names.

[lastzero]

Thank you for your feedback! Much appreciated. If you look at our search filter overview, you'll notice that we use " to quote strings with spaces:

• https://photoprism.app/kb/search-filters

So to search for " you would have to escape it in some way, which many users probably wouldn't do (instrictively). If you use the character for file and folder names, you would also have to replace or escape it somehow. Thus we tried to avoid it as much as possible.

Please note that Unicode search, normalization, and comparison are generally complex topics. While I'm happy to provide you with what I hope is a quick solution, we can't make this a full-time project in the next few weeks to cover all the special cases if the underlying database doesn't already cover them (which needs to be tested before we draw conclusions and develop new functionality).

[lastzero]

A good starting point to learn more about Unicode equivalence in search is the Wikipedia article:

• https://en.wikipedia.org/wiki/Unicode_equivalence