Combine multiple lines until character sequence

Question

Trying to combine multiple lines if it doesn't have an end of message string. Would also like to remove the prefix on all but the first line.

Input:

b929e3e1-29db this is a long message that
abc074df-48a1 is continued onto multiple
dfd4c683-ab48 lines#EOM
ff513a72-570d this is a short message#EOM
980d10aa-bbed another short message#EOM

Output:

b929e3e1-29db this is a long message that is continued onto multiple lines
ff513a72-570d this is a short message
980d10aa-bbed another short message

Ideally using common tools available via command line. (ex: awk, sed)

Answer 1

With GNU awk¹

gawk -vRS='#EOM\n' '{gsub(/\n[[:xdigit:]-]+/,"");} {$1=$1} 1' file.txt

• set the record separator to #EOM followed by a newline
• remove sequences of hex digits and hyphens preceded by newlines; this retains the initial sequence of each record since its newline was eaten by the previous RS
• re-assign a field $1=$1 to force the record to be re-evaluated with default output separators
• print the record (because 1 is true)

Testing

$ gawk -vRS='#EOM\n' '{gsub(/\n[[:xdigit:]-]+/,"");} {$1=$1} 1' file.txt
b929e3e1-29db this is a long message that is continued onto multiple lines
ff513a72-570d this is a short message
980d10aa-bbed another short message

¹. it doesn't appear to work with mawk, for reasons I don't understand (even after replacing + by \{1,\} and various other tweaks)