Login | Register
My pages Projects Community openCollabNet

Discussions > SCons Development (OBSOLETE) > Suggestion for testing WAS: REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner)

Project highlights:

14 Nov 2017: Release 3.0.1 is now available at the download page.

18 Sep 2017: Release 3.0.0 is now available at the download page.

03 Nov 2016: Release 2.5.1 is now available at the download page.

scons
Discussion topic

Back to topic list

Suggestion for testing WAS: REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner)

Author mortenelo
Full name Morten Elo Petersen
Date 2007-03-07 01:01:20 PST
Message Hi,

Just a thought or friendly suggestion!

I am by no means confident with the SCons test suite, but still I would
like to provide a suggestion.

It seems to me that bugs/problems sometimes creeps in or remains
undetected, due to this or that developer not being able to test this or
that tool.

I have had some success in using stubs for the tools for which I do
builders. The reasoning being that when testing my builders the primary
concern is to test that SCons provides the correct files and flags. Of
course at some point I also need to verify that the tools is properly
called, but by using a stub anyone can run the test, even if the tool is
not present on the system.

The stubs are just simple scripts which emulates the actual tools with
respect to accepting input and generating some sort of output. One
approach could be to write the tool command line to the output file, and
then just test that the command line was as expected by inspecting the
file. As all my builders rely on a a construction variable to point at
the binary I can simply replace the actual binary with the stub when
doing my testing.

In my own case my builders silently assume that the tools are present on
the system, without doing any detection, which is a simplification
compared to the test requirements for scons. Also I am aware that
emitters and scanners might make things somewhat more complicated.

regards
Morten

> Gary, I decided to put Dmitry's test in a new test/Intel subdirectory.
> I don't have the Intel compiler installed so I couldn't test
> it directly.
> It looks pretty simple, but I'd appreciate it if you could validate it
> and let me know if it needs a follow-on patch.

« Previous message in topic | 6 of 8 | Next message in topic »

Messages

Show all messages in topic

REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) stevenknight Steven Knight 2007-03-06 11:32:38 PST
     RE: [scons-dev] REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) Sohail Somani <s dot somani at fincad dot com> Sohail Somani <s dot somani at fincad dot com> 2007-03-06 12:30:54 PST
     Re: [scons-dev] REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) garyo Gary Oberbrunner 2007-03-06 13:32:29 PST
         Re: [scons-dev] REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) stevenknight Steven Knight 2007-03-06 13:44:04 PST
             Re: [scons-dev] REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) garyo Gary Oberbrunner 2007-03-06 13:58:12 PST
     Suggestion for testing WAS: REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) mortenelo Morten Elo Petersen 2007-03-07 01:01:20 PST
         Re: [scons-dev] Suggestion for testing WAS: REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) stevenknight Steven Knight 2007-03-07 05:24:09 PST
             Re: [scons-dev] Suggestion for testing WAS: REVIEW: use smart C/C++ linking in the Intel tool chain (Dmitry Girgorenko and Gary Oberbrunner) garyo Gary Oberbrunner 2007-03-07 06:36:04 PST
Messages per page: