Hello,
When I try to script wget, I have to cope with a few problems :
- verbose output is ... verbose, so it is not the way to parse
- -nv output is ... to silent and not easy to parse, as:
- there is no messages about host/domain not founds
- http errors are splitted unto 2 lines (yes, I know grep -B1)
- when using -c, already completed files results in "Refusing to
truncate" message.
So, what I would in an parseable wget output, eg by a
"--scripteable-output" option will be:
- all outputs lines will uses semicolumn (;) separated fields
- first field will have a parseable error message, such as
"HTTP 403", or "UNKNOWN HOST" or "SUCCESS"
- second field will show the file/domain/machine which causes error
- third field will be free, eg help message such as "following File
was not found: http://my.domain/file.html"; or "downloaded n Bytes"
-no bloodly progress bar or incremental downloaded size :)
Also, there is a non consistent exit status:
- if you specify a non-existent file/domain as last url, status is 1
- otherwise, status is 0
Wy wget version is GNU Wget 1.9.1 (and prefer to not change, causes it
is used by apt-methods...)
--
Thomas Harding
