*djvm* problem parsing wildcard argument

[Fooby]

I am trying to create a multi-page djvu from single-page ones with djvm in a bash script. I tried using the * wildcard:

djvm -c "dir1/dir2/My.djvu" "dir1/dir2/djvus/*.djvu"

but I get:

error: *** [1-11711] Failed to open '*.djvu': No such file or directory.

BTW I tried this with and without quotes.

So I instead first created the output file My.djvu using the first file in dir1/dir2/djvus:

d="dir1/dir2/djvus"
f=$(ls "$d" | head -1) # first file in directory
djvm -c "dir1/dir2/My.djvu" "${d}/${f}" # create djvu with first page

And then inserted each page one by one:

for f in dir1/dir2/djvus/*.djvu
do
    # filename.ext split from full path
    p="${f##*/}"
    echo "${p}-->My.djvu"
    djvm -i "dir1/dir2/My.djvu" "${f}" # insert pages one by one
done

And finally deleted the first page which is now a duplicate:

djvm -d "dir1/dir2/My.djvu" 1 # delete first page which is a duplicate

This works OK but it is ridiculously slow because the file gets rewritten from scratch each time.

What is the proper syntax needed to get the wildcard correctly parsed?

[Fooby]

OK, I found what I can only consider a workaround, not really a solution, which IMO would require some modification of the design of this tool.

f=$(echo dir1/dir2/djvus/*.djvu)
djvm -c "dir1/dir2/My.djvu" $f

This only works if there are no spaces in directory and file names.

I created an 862MB multi-page djvu from 1,040 single djvus and it took over 3 hours using the loop method. It took less than a minute using this kludge. But hey, if it works…