Add regress for cross-origin HTTP redirection

Link http regress test with -lz for content-encoding gzip support.
Noticed by anton@

Skip all tests that need internet connection unless they have it.
Run reachability ping only once. Fix clean files and always run
make obj. Do not suppress messages from redirectchain test.

The redirectchain test actually works again. The test servr fixed the
redirect location to be https.

Add RCS markers

The HTTP endpoint used for the redirect chain started to unconditionally
respond with the HTTP protocol in the response location header although
HTTPS is being used in the corresponding request. Flag the test as
expected to fail for now.

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Link http regress test with -lz for content-encoding gzip support.
Noticed by anton@

Skip all tests that need internet connection unless they have it.
Run reachability ping only once. Fix clean files and always run
make obj. Do not suppress messages from redirectchain test.

The redirectchain test actually works again. The test servr fixed the
redirect location to be https.

Add RCS markers

The HTTP endpoint used for the redirect chain started to unconditionally
respond with the HTTP protocol in the response location header although
HTTPS is being used in the corresponding request. Flag the test as
expected to fail for now.

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Skip all tests that need internet connection unless they have it.
Run reachability ping only once. Fix clean files and always run
make obj. Do not suppress messages from redirectchain test.

The redirectchain test actually works again. The test servr fixed the
redirect location to be https.

Add RCS markers

The HTTP endpoint used for the redirect chain started to unconditionally
respond with the HTTP protocol in the response location header although
HTTPS is being used in the corresponding request. Flag the test as
expected to fail for now.

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

The redirectchain test actually works again. The test servr fixed the
redirect location to be https.

Add RCS markers

The HTTP endpoint used for the redirect chain started to unconditionally
respond with the HTTP protocol in the response location header although
HTTPS is being used in the corresponding request. Flag the test as
expected to fail for now.

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Add RCS markers

The HTTP endpoint used for the redirect chain started to unconditionally
respond with the HTTP protocol in the response location header although
HTTPS is being used in the corresponding request. Flag the test as
expected to fail for now.

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Add a test for the redirect loop limit. It would be great if someone
rewrote these tests to use a local HTTP server instead of some random
external ones.

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Include encoding.c

Thanks patrick@

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Automated regress cannot connect to servers in internet. Check
whether jigsaw.w3.org is reachable and skip rpki client http test
if not.
OK claudio@

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Add a minimal regress test for the http client code.
This currently uses some external website to do redirect test and
to check both regular and chunked downloads.
Only for libressl because you can't mix openssl 1.1 and libtls on OpenBSD.

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno

Rework the rpki-client test to also run against OpenSSL 1.1 if the
package is installed. This ensures that the part of rpki-client that
uses libcrypto works as expected with both LibreSSL and OpenSSL.
This will hopefully also help reduce the friction of cutting the
portable release.

The OpenSSL code reaches around into libcrypto for the LibreSSL
specific functions that are provided in -portable. A tiny bit of
glue in test-mft.c is also needed.

Positive feedback from benno