Before this addition you would only have one chance to fill in the Github credentials.
If they where wrong the script would fail on a lot of the remaining steps, you would need to run the script and you might even get issues because of already used commands.
This commit adds the response from Github to check if the credentials are valid are not.
In case they're invalid we'll ask them again (untill they're correct) and we'll show a warning to the user that they where invalid or that the user does not have enough rights on the Github repository.
If they succeed we'll clone the enterprise repository and continue.
This will automatically fill in the XMLRPC-port that is configured at the top of the script.
This was not passed on in a variable so if the user would change the port at the top it wouldn't change in the configuration file. This is now fixed!
Added the "--depth 1" to the command so it only retrieves the latest version without all the repository history, which makes the download quite a bit quicker, and uses less space in disk, making it suitable for a production environment.
- Integer overflow test does not worked properly and is not reliable
indicator of host architecture, it replaced with "getconf LONG_BIT"
that was tested and worked properly on Ubuntu 14.04.
now wkhtml2pdf downloads are correctly adapted to x64 and x32 hosts.
- Automatically choose between wkhtml2pdf x64 and x32 versions,
according of host architecture.
- Replaced "dpkg -i" with "gdebi --n", in order to have installed
all wkhtml2pdf dependencies automatically, if any of them is missing
from the system.
- Make shortcuts in /usr/bin directory, instead of copying actual binary
files that are part of installed wkthml2pdf package.
- Added .gitignore in order to prevent garbage file upload to the
repository.
- In order to automate installation, added -y option to "apt-get
install" command, as without -y the command enters interactive
mode with a question below:
"Do you want to continue? [Y/n]"