Code to configure the Apache web server appears in multiple places, i.e.,
While I like the idea of a "reference implementation" for setup, I really do not like the appearance of this code in prepare-company-database.sh because it is reasonable to assume that the script will be utilized more than once, i.e. to establish records for more than one company. When that is done, any additional, manual changes made to the Apache configuration are lost by overwriting the configuration file.
That is to say, after initial installation, after I have configured the web server once using configure_apache.sh to set the permissions on the LedgerSMB application directory, I manually edit the resulting, copied ledgersmb-httpd.conf so as to include the local network IP addresses from which users actually run and SSL configuration. I do not want to subsequently have to redo that manual configuration every time I run prepare-company-database.sh. Even if I am not creating records for more than one company, during testing I am frequently running
delete-company-database.sh && prepare-company-database.sh
Having to re-configure the web server each time got old pretty fast.
If you agree, the attached patch against trunk removes the "sed ..." line.