#323 cuesheet parser can't handle BOM marker

[Josh Coalson]

--- Artyom Karpenko <****@gmail.com> wrote:

> Is there any way for flac or metaflac to support input from files
> with Unicode Byte Order Marks (BOM)?
> In command:
> metaflac --remove-all-tags --set-tag-from-file="CUESHEET=CDImage.cue"
> --set-tag-from-file="LOGFILE=CDImage.log" CDImage.flac
>
> If "CDImage.cue" contains Unicode characters, most text editors
> append BOM at the 1st line of this file.
>
> And metaflac does not deal with it. It simply put BOM to the flac as
> part of the tag value. So the CUESHEET tag appears broken.

[Nobody/Anonymous]

Logged In: NO

artyomk3d@gmail.com
Example cuesheet:
First 3 symbols are automatically added by notepad to any Utf-8 text file.
---
я╗┐REM GENRE Rock
REM DATE 1997
REM DISCID A30E1A0C
REM COMMENT "ExactAudioCopy v0.95b4"
PERFORMER "U2"
TITLE "Pop (1997 Island CIDU210 524 334-2)"
FILE "CDImage.wav" WAVE
TRACK 01 AUDIO
TITLE "Discoth├иque"
INDEX 01 00:00:00
TRACK 02 AUDIO
TITLE "Do You Feel Loved"
INDEX 01 05:18:50
TRACK 03 AUDIO
TITLE "Mofo"
INDEX 01 10:26:05
TRACK 04 AUDIO
TITLE "If God Will Send His Angels"
INDEX 00 16:13:10
INDEX 01 16:15:15
TRACK 05 AUDIO
TITLE "Staring At The Sun"
INDEX 01 21:37:50
TRACK 06 AUDIO
TITLE "Last Night On Earth"
INDEX 01 26:14:35
TRACK 07 AUDIO
TITLE "Gone"
INDEX 01 31:00:10
TRACK 08 AUDIO
TITLE "Miami"
INDEX 01 35:26:67
TRACK 09 AUDIO
TITLE "The Playboy Mansion"
INDEX 01 40:19:62
TRACK 10 AUDIO
TITLE "If You Wear That Velvet Dress"
INDEX 01 45:00:50
TRACK 11 AUDIO
TITLE "Please"
INDEX 01 50:15:22
TRACK 12 AUDIO
TITLE "Wake Up Dead Man"
INDEX 01 55:17:65
---

[gharris999]

Logged In: YES
user_id=1734443
Originator: NO

Why is this a problem? Are you saying that some software (other than flac and metaflac) has trouble reading comment[x]=CUESHEET content if it contains a BOM? None of the software that I use has any problem with this. I regularly store a BOM in my embedded metadata cuesheets.

If your cuesheet is UTF8 encoded with a BOM, then embed the cuesheet with:

metaflac --no-utf8-convert "--set-tag-from-file=CUESHEET=cuesheet file.cue"

and, as you say, the BOM will be embedded in the flac along with the correctly encoded UTF8 data.

To Extract, I use:

metaflac --no-utf8-convert --show-tag=CUESHEET flacfile.flac >"some cuesheet.cue"

Then, you need to strip out the first 9 bytes of the file (the "cuesheet=" part) and you'll be left with a properly UTF8 encoded cuesheet with the correct BOM.

Personally I can see the utility of an enhancement to metaflac: a "--export-tag-to=TAGNAME=filename" option that would be the "data out" partner to the existing "--set-tag-from-file=TAGNAME=filename" option. This would obviate the need to strip out the tagname header from the output file.

That's what I'd like to see, anyway.

[gharris999]

Logged In: YES
user_id=1734443
Originator: NO

Actually, I suppose that metaflac enhancement option ought to be:

--export-tag-to-file=TAGNAME=filename

[gharris999]

Logged In: YES
user_id=1734443
Originator: NO

OK, so metaflac --no-utf8-convert "--import-tags-from=somefile.txt" can't handle somfile.txt if it has a BOM header. Not a big deal, but "--no-utf8-convert" ought to be a clue that a BOM may be present and ought to be skipped past.

[Erik]

Closing old out-of-date issue.