clean up IPTC Description/Caption field

[Phil Harvey]

If you are passing special characters in file names on the command line, you must set the -charset filename=XXX option to set the character set to whatever you are using.  See here for more information.

- Phil

[frereroy]

Thanks. Am using the -L switch (for Windows Latin1).

As the job progresses I see "funny" characters in the paths that contain accents but the filenames themselves, which do not contain accented characters, are treated correctly.

[Phil Harvey]

-L is for tag values.  The -charset filename=XXX option is for file/directory names.  There is some overlap because the FileName and Directory tags use both options.

Is the job working aside from seeing funny characters?  An where are you seeing them?

- Phil

[frereroy]

I am now using the following command

exiftool -ext JPG  -charset filename=Latin -use MWG "-Description<${Description;s/\s+$// or $_=undef}" "-Caption-Abstract<${Caption-Abstract;s/\s+$// or $_=undef}" -overwrite_original -r /Phototheque

The job is working well but the accented directory names show as gobbledy-gook as the lines scroll through the command line box.

[Phil Harvey]

If you are talking about the file name after the "====" when processing multiple files, then yes.  There is currently no character translation for the informational messages.  I should maybe look into doing this.

- Phil

[frereroy]

I do not see the "===="

The funny thing is that if I write ".... -r /Phototheque" then all embedded directories with accents in their names will be processed albeit showing the odd characters on the screen while processing but if the top level directory contains an accent like ".... -r /Photothèque" then I get the error  "File not found". and the underlying directories are not parsed.

[Phil Harvey]

That's not funny at all.  Specifying file names with special characters is problematic as I said.  It can be done, but you must get your character sets correct.

- Phil

[frereroy]

Code Select Expand

exiftool -ext JPG -use MWG "-Description<${Description;s/\s+$// or $_=undef}" "-Caption-Abstract<${Caption-Abstract;s/\s+$// or $_=undef}" -overwrite_original -common_args -charset filename=cp1252 -r /Photothèque


Now works. i.e. the top level directory is read correctly with it's accent and parsing the directories within works inspite of showing "odd" characters on the screen instead of the accented ones.

I was rather hoping that the -common_args tag that I added would correct this.