Allow skip files that fail to get the contents during downloading logs.
Will show in command message to let downloader knows what file we skips,
so downloader can try to download it manually if they prefer.
Change-Id: Ic87e1e2f1ef656b4d0615fa6fb68a1ed9067a956
- Do not insert an extra slash into the download URL, this breaks on
some storage providers.
- Do not try to uncompress *.xz files, libcurl cannot handle that
compression, also these are mostly journal files that are compressed
for a good reason.
- Mention the download directory again at the end, avoids needing to
scroll through umpteen log messages in order to find out the name of
the tmpdir generated at the start of the script.
Change-Id: I8eb674121a9808dfc8edfc14b8c3525a77b518ab
This is an alternative to I98c80f657f38c5e1ed5f28e5d36988a3429ad1f8
which does not modify the upload script, but rather queries the API
and manifest for what to download.
The script is a hybrid of python and bash to not implement json API
parsing badly in bash, but not replicate curl badly in python either.
The script sanity checks for dependencies, which are considered pretty
standard for any developer who would be interested in downloading logs
like this.
The role writes out the script with the correct build values coded
into it, so a potential user just has to run it without any arguments
or modification.
Change-Id: Ic33732adbfd3210191bf4976c3ee316cfc50568e