Rebase to openresty 1.19.9.1
- Mar 24, 2022
-
-
Francesco Giacomini authored
-
- Dec 01, 2021
-
-
Francesco Giacomini authored
* Explicitly install build-time deps needed for this package * Take the httpg patch from this repo
-
Francesco Giacomini authored
The script can be used both for the devcontainer and for the continuous integration.
-
Francesco Giacomini authored
-
- Nov 16, 2021
-
-
Francesco Giacomini authored
The debug spec file is not used, but let's keep it aligned with the other one.
-
- Nov 15, 2021
-
-
Francesco Giacomini authored
-
- Nov 12, 2021
-
-
Francesco Giacomini authored
* Start from the upstream SPEC file. * Use the %{name} macro instead of explicitly repeating openresty-voms * Use system zlib, openssl and pcre dependencies and do not specify minimum versions * Remove unused macros oroprefix, {zlib,pcre,openssl}_prefix * The voms_module_prefix macro is assumed * Mv-ing openresty binary in openresty-voms seems irrelevant * Don't consider "tapset" files, they don't exist (any more) * For debug builds use -Og * Review the packaging script - Minimize the list of build-time packages to install - Set the voms_module_prefix macro in the .rpmmacros file. The prefix is on the command line, with a default - Build only the optimized RPM * Start docker:dind with --tls=false to avoid timeouts during CI Issue #17
-
- Nov 11, 2021
-
-
Francesco Giacomini authored
-
Francesco Giacomini authored
CentOS provides gcc 4.8.5, which has a decent support for C++14 features, enabled with -std=c++1y. In this way we don't need to bring in the devtoolset software collection.
-
Francesco Giacomini authored
Conflict is set against any version of openresty.
-
Francesco Giacomini authored
-
Francesco Giacomini authored
In particulare the one in certs shows concrete examples on how to generate VOMS proxies
-
Francesco Giacomini authored
The proxies had a wrong VO in the FQANs and couldn't be validated any more by the VOMS lib. See commit https://github.com/italiangrid/voms/commit/feeaed encoding.t was also affected, but didn't require any change, just a correctly generated proxy.
-
Francesco Giacomini authored
Apparently the bug mentioned in the (removed) comment has been fixed. But this showed a bug in our test: http vs HTTP in the pattern to match in the log. Issue #17
-