Quality assurance

From GNUstepWiki

(Difference between revisions)
Revision as of 18:00, 24 February 2005
Stefan Urbanek (Talk | contribs)

← Previous diff
Revision as of 07:45, 29 October 2006
Rfm (Talk | contribs)
Introduction
Next diff →
Line 1: Line 1:
== Introduction == == Introduction ==
-Sometimes the GNUstep does not work as expected or can not be built at all. If the actual code which gets committed (which IMHO is really a basline requirement) was tested then this should not happen. Things which break the entire build cycle should be taken more seriously.+Sometimes the GNUstep does not work as expected or can not be built at all. If the actual code which gets committed (which IMHO is really a basline requirement) was tested then this should not happen. Unfortunately there is no waay to guarantee this as it's always possible to make errors in comitting. However, there are two things which can help with this.
 + 
 +As a standard procedure developers should build the code and run the testsuites (http://svn.gna.org/viewcvs/gnustep/tests/testsuite/trunk/) before comitting changes.
 + 
 +Clean builds are run on a nightly basis and the results are sent to the gnustep developers mailing list. It is important that people volunteer to have their machines participate in this process so that developers can be made aware os system dependent problems as soon as possible.
 +See http://wiki.gnustep.org/index.php/Developer_FAQ#How_can_I_take_part_with_a_GNUstep_autobuilder_for_the_testfarm.3F
[[User:Aperez|Aperez]]: '' I'd also like to set this up for win32, since that seems to be the thing that breaks fairly often as other patches go into -core. '' [[User:Aperez|Aperez]]: '' I'd also like to set this up for win32, since that seems to be the thing that breaks fairly often as other patches go into -core. ''

Revision as of 07:45, 29 October 2006

Introduction

Sometimes the GNUstep does not work as expected or can not be built at all. If the actual code which gets committed (which IMHO is really a basline requirement) was tested then this should not happen. Unfortunately there is no waay to guarantee this as it's always possible to make errors in comitting. However, there are two things which can help with this.

As a standard procedure developers should build the code and run the testsuites (http://svn.gna.org/viewcvs/gnustep/tests/testsuite/trunk/) before comitting changes.

Clean builds are run on a nightly basis and the results are sent to the gnustep developers mailing list. It is important that people volunteer to have their machines participate in this process so that developers can be made aware os system dependent problems as soon as possible. See http://wiki.gnustep.org/index.php/Developer_FAQ#How_can_I_take_part_with_a_GNUstep_autobuilder_for_the_testfarm.3F

Aperez: I'd also like to set this up for win32, since that seems to be the thing that breaks fairly often as other patches go into -core.

Compilability test

As it is frustrating to check out CVS and have it fail to build for such a trivial reason.

Some sort of automated QA mechanism is necessary, where, upon a new CVS check-in to -base, for example, it is be checked out by an automated QA system and compiled on Linux/x86 as a baseline test. The results of the compilation would be placed on a webpage. If compilation breakage to any part of -core happened, a message would be sent to -dev notifying the core team that this had happpened, the offending file, and the version of the file in CVS at which the breakage occurred.

Idea of a primitive test script for compilability:

cd test_directory
rm -rf core GNUstep-test-root
cvs co core
cd core
./build-all $test_directory/GNUstep-test-root &> test-log.txt