list of disallowed special characters for filenames is not accurate

[matthew-a-dunlap]

Working on CORE2 which uploads to dataverse, I'm writing code to prevent certain characters in filenames.

Testing Dataverse's restrictions, it reports File Name cannot contain any of the following characters: / : * ? " < > | ; # .

This info is incorrect for at least two reasons:

1. This filename still errors even though none of those characters are present: !$%&’()+,-=@[\]^_{}~. It seems this is because the \ character is prevented as well.
2. the . character is not actually prevented by Dataverse as far as I can tell, even if its used multiple times.

It would be useful for the user to have a more accurate error message. Also would be useful if this was in the documentation (I may have missed it). This happens on 5.3 and 5.10.1

[shlake]

Working on this, but have found another problem. See this #9080

As noted above \ isn't "allowed" in a filename, but the Dataverse software changes the filename with \ , but it doesn't generate an error. This filename citations\files.txt is changed to files.txt without an error.

[poikilotherm]

Related:

• directoryLabel leading '.' char for hidden directories is ignored (removed) #8807

[shlake]

@matthew-a-dunlap I'm checking on non-valid characters, but I have figured out that the . in the error list, is just the end of the sentence . it does not mean that "period" is an invalid character, but it does need to be removed to avoid confusion.

The error message is coming from this file: WEB-INF/classes/ValidationMessages.properties

[ErykKul]

Patter for the "label" in the code: https://github.com/IQSS/dataverse/blob/develop/src/main/java/edu/harvard/iq/dataverse/FileMetadata.java#L72
regexp="^[^:<>;#/\"\\*\\|\\?\\\\]*$"

[ErykKul]

Directory name validator: https://github.com/IQSS/dataverse/blob/develop/src/main/java/edu/harvard/iq/dataverse/FileDirectoryNameValidator.java#L32

String validCharacters = "[\\w\\\\/. -]+";