-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Eric -- Definitely! I completely agree. This patch wasn't intended to
address the bug in the discuss-list. I think it's useful regardless of
the solution to that problem -- it just adds a sanity check on input,
which seems useful.
As long as the fix to that problem doesn't change the API to the filter
constructors (which seems unlikely for backwards compatibility), this
sanity check will be correct and a good hint to users. Unless I'm
misunderstanding something.
- -Dan
Eric Blossom wrote:
> Dan,
>
> The reason it was backed out was that it has "make check" fail.
> Johnathan didn't back it out because it wasn't right or wrong,
> he backed it out because it causes "make check" to fail.
>
> The fix can go back in when the QA code has been updated.
> The fix and the updated QA code should come as a single patch that.
>
> Eric
>
>
> On Tue, Dec 11, 2007 at 08:31:46PM -0800, Dan Halperin wrote:
> Attached is a patch for gnuradio-core/src/lib/filter/gri_iir.h that
> includes a sanity check on the input. See the forwarded message.
>
> I'm sure the error message could be better and instead of exit(0) an
> exception could be thrown or something -- if you would please clean it
> up for me I'll know how to do it correctly the next time :).
>
> -Dan
- --- gri_iir.h 2007-12-11 20:28:17.000000000 -0800
+++ tmp 2007-12-11 20:27:17.000000000 -0800
@@ -90,6 +90,11 @@
const std::vector<tap_type> &fbtaps) throw
(std::invalid_argument)
{
+ if (fbtaps.size() > 0 && fbtaps[0] != 0)
+ {
+ fprintf(stderr, "Are you using fbtaps correctly? The first
value should always be zero!\n");
+ exit(0);
+ }
d_latest_n = 0;
d_latest_m = 0;
Date: Tue, 11 Dec 2007 20:16:25 -0800
From: Dan Halperin <[EMAIL PROTECTED]>
Subject: Re: [Discuss-gnuradio] make check errors
To: Johnathan Corgan <[EMAIL PROTECTED]>
Cc: gnuradio mailing list <[EMAIL PROTECTED]>
Sender: [EMAIL PROTECTED]
>>
- -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
>>
Johnathan Corgan wrote:
>>> On 12/11/07, Johnathan Corgan <[EMAIL PROTECTED]> wrote:
>>>
>>>> I see this as well now. It can be attributed to a check-in on the trunk
>>>> 2 days ago that Martin did to fix an iir bug (r7089). I suspect the iir
>>>> QA needs to be updated.
>>> The changeset for r7089 has been reverted on the trunk.
I looked through that code and the formulas in the header file and did
the QA by hand and it seems that the reverted version is/was correct.
Perhaps there was an error forgetting to prepend a 0 into the fb_taps
that made someone think there was a bug?
>>
- - -Dan
- -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>>
iD8DBQFHX2CZy9GYuuMoUJ4RAjKgAJ4l5cQ6B8vJUqyEbuMvZhfoQZRbSQCdFJTG
QiFBO/PjvRX1y96QvGT43Vs=
=x9HD
- -----END PGP SIGNATURE-----
>>
>>
_______________________________________________
Discuss-gnuradio mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________
Patch-gnuradio mailing list
Patch-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/patch-gnuradio
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFHX3ZHy9GYuuMoUJ4RAj7wAJ9fRqNzAWTmNluuxTKk1MOPwwrNxQCggKjc
rt3m3MdM98dB9rmGnuWJH2o=
=IdAH
-----END PGP SIGNATURE-----
_______________________________________________
Patch-gnuradio mailing list
Patch-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/patch-gnuradio
