This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: EXEEXT settable in generated buildtree, via CDL ??
- From: Bart Veer <bartv at ecoscentric dot com>
- To: shimple0 at yahoo dot com
- Cc: ecos-discuss at ecos dot sourceware dot org
- Date: Fri, 26 Nov 2004 16:28:23 +0000 (GMT)
- Subject: Re: [ECOS] EXEEXT settable in generated buildtree, via CDL ??
- References: <20041126055157.48331.qmail@web52704.mail.yahoo.com>
>>>>> "sandeep" == sandeep <shimple0@yahoo.com> writes:
sandeep> to set the extension of test images, one can edit
sandeep> toplevel makefile of buildtree and set EXEEXT to desired
sandeep> extension.
sandeep> many months back, i had made a minor addition in
sandeep> pkgconf/rules.mak
>> export EXEEXT := $(EXEEXT)
sandeep> could this be a trivial patch, if useful??
sandeep> using which i could avoid the need to edit generated
sandeep> makefile. i just needed to
sandeep> - either "export EXEEXT=.xyz" in execution environment
sandeep> - or "EXEEXT=.xyz make tests"
sandeep> then i had searched in documents/internet to find a CDL
sandeep> way for this, but couldn't locate. also i had tried some
sandeep> calculated guesses on cdl names for this but no success.
The idea of having suffixes for test executables was considered very
briefly in the early days of the eCos design, and rejected. They are
more trouble than they are worth. All the information you need about
an executable can be found in the executable itself, e.g.:
$ file tm_basic
tm_basic: ELF 32-bit MSB executable, Motorola 68020, version 1 (SYSV), statically linked, not stripped
The EXEEXT support in rules.mak probably dates back from the days of
gcc 2.95 when gcc on Windows always appended a .exe suffix to an
executable, irrespective of whether that executable was a Windows one
or for an embedded target. It causes all kinds of problems, e.g. test
farm code had to work differently on Linux vs. Windows. Pretty much
everybody involved with gcc for embedded systems opposed that
behaviour and current gcc behaves sensibly. The EXEEXT support in
rules.mak should probably be expunged.
Of course if an application developer wants to use some kind of
suffix for executables, that is fine. eCos does not mandate how
applications get built.
Bart
--
Bart Veer eCos Configuration Architect
http://www.ecoscentric.com/ The eCos and RedBoot experts
--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss