On Thu, Sep 10, 2026 at 3:21 AM Joe Orton <[email protected]> wrote: > > On Wed, Sep 09, 2026 at 01:51:54PM +0000, Cornel Isbiceanu via dev wrote: > > Hi all, > > > > I'd like to propose a small addition to mod_env: a new SetEnvFromFile > > directive that reads internal environment variables from a file of > > name=value lines at configuration time, storing them in the same > > per-directory table used by SetEnv. > > > > PR (against trunk): > > apache/httpd#750<https://github.com/apache/httpd/pull/750> > > > > Motivation: it lets operators keep a set of environment values in a > > separate, easily-generated/managed file (e.g. from deployment tooling) > > instead of inlining many SetEnv lines in the server config. Values flow to > > r->subprocess_env exactly like SetEnv, so they reach CGI, SSI, logging, > > proxy backends, etc. unchanged. Because both write the same table, > > precedence follows configuration order (last writer wins), matching > > existing SetEnv behaviour. > > > > The file format is one name=value per line; blank lines and # comments > > are ignored, whitespace is stripped, and relative paths resolve > > against ServerRoot. It's read once at config parse time. > > Thanks for filing the PR! Seems like a nice feature. > > It's probably worth documenting that backslash continuation is respected > in the files. You can credit yourself in the changes-entry file. > Otherwise LGTM. > > Also if you don't mind burning some more tokens, a pyhttpd-based test > would be preferred over the pytest_suite one (yes we should document > that) - I just pushed another test to test/modules/metadata which would > be the natural home for mod_env tests too.
Any issue with it allowing unintended files via .htaccess that have some content matching the format ? Seems like the kind of thing we'd get reports on. Maybe a new AllowOverride type?
