The noted test failures in http.test are not based on broken nsd
behavior, but are false positives due to the move from low-level I/O to
ns_http in the regression test (makes test with http und https uniform).
the ns_http encoding issues still have to be addressed, but the
following change make these FAILED results go away and explain the
change programmatically.
https://bitbucket.org/naviserver/naviserver/commits/bc4a94384a6f2d74587a2eded54706489fd694f7
all the best
-g
On 15.07.19 10:46, Andrew Piskorski wrote:
Using the latest NaviServer head (last commit 2019-07-09) on Linux
(Ubuntu 16.04.2 LTS) and Tcl 8.6.9 (built from source), I'm getting
'make test' failures that I don't understand, primarily in http.test.
Can anyone advice, please? Brief failure messages are below, full
test output is attached.
Tests ended at Mon Jul 15 04:23:22 EDT 2019
all.tcl: Total 1391 Passed 1380 Skipped 8 Failed 3
Sourced 69 Test Files.
Files with failing tests: http.test ns_server.test
==== http-3.3new limits: too many headers FAILED
==== Contents of test case:
nstest::http -http 1.1 -getbody 1 -setheaders [split [string repeat xy 1024]
""] GET /limits
---- Result was:
ns_http failed: recv failed
---- Result should have been (exact matching):
414
==== http-3.3new FAILED
==== http-5.5 check encoding ns_conn content POST, binary FAILED
==== Contents of test case:
set string "Testing <äöüß>"
nstest::http -getbody 1 -setheaders {content-type text/html} POST /post
$string
---- Result was:
200 {utf-8 <text/html> AÄATesting <äöüÃ>ZÜZ}
---- Result should have been (exact matching):
200 {utf-8 <text/html> AÄATesting <äöüß>ZÜZ}
==== http-5.5 FAILED
==== ns_server-2.5 basic operation FAILED
---- Result was:
10
---- Result should have been (exact matching):
9
==== ns_server-2.5 FAILED
_______________________________________________
naviserver-devel mailing list
naviserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/naviserver-devel
