USE flags for testing: /
objective: `emerge mapserver`
test system: x86-macos
Created attachment 133383 [details, diff]
Patch to webapp-config 1.50.16-r2 ebuild to make it work with prefixed portage
This ebuild needs a file in $FILES_DIR to be present which will be attached separately.
Created attachment 133384 [details, diff]
Patch to webapp-config 1.50.16 itself to work with the prefixed portage
There was a notification I got regarding the sandbo. I got it when adding new applications to my apache2, but I'm unable to reproduce it now. It maybe that something in webapp-config needs further tuning.
while os.environ['EPREFIX'] might work, the proper way to do this is @GENTOO_PORTAGE_EPREFIX@ in the patch and eprefixify in the ebuild.
Created attachment 133975 [details, diff]
Updated patch to webapp-config 1.50.16 itself to work with the prefixed portage
- changed to use @GENTOO_PORTAGE_EPREFIX@
- appended patch for webapp-cleaner
Since the sandbox hasn't been ported to Darwin, I can't test how the sandbox related cod works.
Created attachment 133976 [details, diff]
Updated patch to webapp-config 1.50.16-r2 ebuild to make it work with prefixed portage
Created attachment 133977 [details, diff]
webapp.eclass patch to make it work in prefixed portage
webapp-config related apps rely on an eclass that I forgot to include in this bug to make the whole thing work.
Some comments for further improver:
webapp-config is called from the command line and the eclass is called and from ebuilds. Both use a common configuration file that has the prefix already included. For that reason the functions called from the ebuilds that rely on functions like `dodir` need some special care, as for those the path needs to be altered. It was further complicated by eclass internal variables (MY_*) that are defined once and then used through successive calls during the webapp ebuilds.
And now the webapp.eclass got updated so changes worked on by Dirk may be outdated. lovely..
sorry .... :(
I have self motivation for getting trac into the prefix tree and that depends on webapp-config. Since webapp.eclass has been changed since your bug and comment #8, someone will have to look at it later. Thanks for your changes, if you have anything newer, please re-open this bug and I will look at it.
webapp-config is in the tree for now. I used your patch in conjunction with a patch from bug 206973, both had about the same content. Thanks!
hmmm, sorry, that was not my intention