On 11/30/2018 11:48 PM, Steve Ellcey wrote:
> On Thu, 2018-11-29 at 22:59 +, Pedro Alves wrote:
>>
>> Yeah, GDB does in its internal board files. Non-remote boards source
>> a file with this:
>>
>> ~
>> 16 # By default, DejaGnu makes the board remote unless the board
>> name
>> 17 # ma
On Thu, 2018-11-29 at 22:59 +, Pedro Alves wrote:
>
> Yeah, GDB does in its internal board files. Non-remote boards source
> a file with this:
>
> ~
> 16 # By default, DejaGnu makes the board remote unless the board
> name
> 17 # matches localhost. Sourcing this script from a board
On 11/28/2018 02:27 AM, Ben Elliston wrote:
> On Tue, Nov 27, 2018 at 06:53:17PM -0600, Jacob Bachmeyer wrote:
>
>> As we see, set_board_info only works if the specified /entry/ has not
>> already been defined.
>
> That is consistent with the manual:
>
> @node set_board_info procedure, [...]
>
Ben Elliston wrote:
On Wed, Nov 28, 2018 at 09:55:15PM -0600, Jacob Bachmeyer wrote:
I believe that your pessimism is premature: if Wikipedia is to be
believed, there was a new release 8.6.9 12 days ago on November 16.
I do not think that we need to worry about Tcl going away anytime
soon. :
On 11/29/18 12:35 AM, Brooks Moses wrote:
> Fair enough -- but having done cross-testing on both GCC (with
> DejaGnu) and glibc (without it), and having also done some substantial
> porting work to get the GCC testing working with our systems, I have
I'll note that GCC testing is much easier th
On Wed, Nov 28, 2018 at 11:35:12PM -0800, Brooks Moses wrote:
> Admittedly, that's because the "figure out which tests to run" and
> "run a test on this kind of machine" functionality is provided by
> our build-farm infrastructure, but I still found it noteworthy how
> separable the pieces were.
On Wed, Nov 28, 2018 at 5:10 PM, Rob Savoye wrote:
> On 11/28/18 2:10 PM, Brooks Moses wrote:
>> Much like the other big lesson of the last 30 years or so: Shipping
>> something that works is far more valuable than waiting to try to
>> achieve perfection. :)
>
> It was worse than that. :-) When
On Wed, Nov 28, 2018 at 09:55:15PM -0600, Jacob Bachmeyer wrote:
> I believe that your pessimism is premature: if Wikipedia is to be
> believed, there was a new release 8.6.9 12 days ago on November 16.
> I do not think that we need to worry about Tcl going away anytime
> soon. :-)
Expect, on th
Rob Savoye wrote:
My daydream is to analyze the current functionality, write a true
design doc, and write a new python based implementation. :-) We need
something that'll last another 30 years, and last I checked, Tcl has no
maintainers...
I believe that your pessimism is premature: if Wi
On 11/27/18 3:46 PM, Steve Ellcey wrote:
Sorry, catching up from being offline while traveling...
> I have a question about the special handling of the 'unix' name in
> dejagnu. I am doing some GCC testing which uses dejagnu and I
> want to create a new baseboard (unix-sysroot) which is just
On 11/28/18 4:26 PM, Jacob Bachmeyer wrote:
> $DEJAGNU and the local init files ($base_dir/site.exp, $objdir/site.exp)
> are loaded independently. It does not help that both local and global
> init files are named (or default to) "site.exp". Documenting this is on
> my TODO list.
The site.exp
On 11/28/18 2:10 PM, Brooks Moses wrote:
> Much like the other big lesson of the last 30 years or so: Shipping
> something that works is far more valuable than waiting to try to
> achieve perfection. :)
It was worse than that. :-) When I started DejaGnu, there were only 7
of us in all of Cygnu
Ben Elliston wrote:
On Wed, Nov 28, 2018 at 04:57:28PM +, Steve Ellcey wrote:
The main reason for not creating a new site.exp file is that the
testing I am doing with dejagnu is running the GCC testsuite and the
GCC 'make check' creates the site.exp file on the fly.
You could set $
On Wed, Nov 28, 2018 at 04:57:28PM +, Steve Ellcey wrote:
> The main reason for not creating a new site.exp file is that the
> testing I am doing with dejagnu is running the GCC testsuite and the
> GCC 'make check' creates the site.exp file on the fly.
You could set $DEJAGNU to your own site.
On Wed, Nov 28, 2018 at 12:50 PM, Ben Elliston wrote:
> On Wed, Nov 28, 2018 at 08:07:36AM -0700, Rob Savoye wrote:
>> I always wanted to fix the lack of design, and also not break any
>> existing testsuites, but after trying to find funding for this for
>> years and failing, never got around to
On Wed, Nov 28, 2018 at 08:07:36AM -0700, Rob Savoye wrote:
> I always wanted to fix the lack of design, and also not break any
> existing testsuites, but after trying to find funding for this for
> years and failing, never got around to it.
That has to be one of the biggst lessons of software
On Wed, 2018-11-28 at 13:38 +1100, Ben Elliston wrote:
> Why not just put this into site.exp?
>
> append LDFLAGS $env(DEJAGNU_UNIX_SYSROOT_FLAGS)
>
> As Jacob indicated, a new baseboard is the wrong hammer.
Just to clarify, when Jacob suggests 'creating a new target',
is that the same as cre
On 11/27/18 7:27 PM, Ben Elliston wrote:
> On Tue, Nov 27, 2018 at 06:53:17PM -0600, Jacob Bachmeyer wrote:
>
>> As we see, set_board_info only works if the specified /entry/ has not
>> already been defined.
>
> That is consistent with the manual:
>
> @node set_board_info procedure, [...]
>
On 2018-11-27 5:46 p.m., Steve Ellcey wrote:
> I have a question about the special handling of the 'unix' name in
> dejagnu. I am doing some GCC testing which uses dejagnu and I
> want to create a new baseboard (unix-sysroot) which is just like
> unix but sets ldflags so I can run GCC with a newl
Hi Steve
> if {[info exists env(DEJAGNU_UNIX_SYSROOT_FLAGS)]} {
> set_board_info ldflags "$env(DEJAGNU_UNIX_SYSROOT_FLAGS)"
> }
Why not just put this into site.exp?
append LDFLAGS $env(DEJAGNU_UNIX_SYSROOT_FLAGS)
As Jacob indicated, a new baseboard is the wrong hammer.
Cheers, Ben
signat
On Tue, Nov 27, 2018 at 06:53:17PM -0600, Jacob Bachmeyer wrote:
> As we see, set_board_info only works if the specified /entry/ has not
> already been defined.
That is consistent with the manual:
@node set_board_info procedure, [...]
This checks if @code{board_info} array's field @emph{entr
Steve Ellcey wrote:
I have a question about the special handling of the 'unix' name in
dejagnu. I am doing some GCC testing which uses dejagnu and I
want to create a new baseboard (unix-sysroot) which is just like
unix but sets ldflags so I can run GCC with a newly built glibc
in a non-standard
22 matches
Mail list logo
