[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: get-package-attribute sometimes returns "?" - ID: 3114991





Date: Thu, 16 Dec 2010 21:56:28 -0700
From: John Doty <jpd@xxxxxxxxx>
Subject: Re: gEDA-user: get-package-attribute sometimes returns "?" -
      ID:     3114991

On Dec 16, 2010, at 9:29 PM, clif@xxxxxxxxxxxxx wrote:

I'm not sure we're on the same page here. That pice of legacy code is
what we want to remove. We don't want any code to reinterpret
attributes. For the short term we would like to change a few
probramatic symbols so we don't have to have special cases in the code.

But the "?" is there to tell you, when looking at the symbol in gschem,
that you *need* to edit the attribute (I use those symbols a lot).

Yes, agreed that is why it is there. However you may not see it unless you click on show inhereted attributes, and because they are inherited you can't delete them. You can only delete the promoted ones after which the inherited ones take over again.

Sometimes I use the blocks without the file name, just a value attribute to get a one liner. Other times I start with one type of block but change it to another, so I can see cases where a particualr attribute becomes superfluous.

I'm puzzled here, because I thought Stuart was the author of both
spice-sdb and the symbols in question, so the "WTF?" is strange.

Yes and it is causeing confusion so I would like to find a better way to handle it.

In any case, it seems to me that the user should get an error message
(from spice-sdb) in this case, as an explicit "?" here means that the
user has neither set the attribute to something useful nor deleted it as
irrelevant.

However the WTF? snipit is allowing it to fail silently.

John Doty              Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd@xxxxxxxxx

Date: Fri, 17 Dec 2010 08:21:45 +0000
From: Peter TB Brett <peter@xxxxxxxxxxxxx>
Subject: Re: gEDA-user: get-package-attribute sometimes returns "?" -
      ID:     3114991

So what you are saying is that you think we should get rid of the "WTF?"
function in the code, but not modify the symbol library?

               Peter

--
Peter Brett <peter@xxxxxxxxxxxxx>
Remote Sensing Research Group
Surrey Space Centre

Date: Fri, 17 Dec 2010 07:52:15 -0700
From: John Doty <jpd@xxxxxxxxx>
Subject: Re: gEDA-user: get-package-attribute sometimes returns "?" -
      ID:     3114991

On Dec 17, 2010, at 1:21 AM, Peter TB Brett wrote:

I think there's nothing wrong with file=? in the symbols. It indicates
that the you have failed to supply a required filename.

Or an optional filename.

(create-file-info-list), the "WTF?" function in spice-sdb, should issue
a diagnostic if the file name is "?".

Perhaps that could be a quick fix but currently it fails silently.

Right now, all that results from
substituting "unknown" is that it silently omits the file, which is
surely wrong behavior. You won't even know there's a problem until you
run SPICE, which has its own troubles with generating diagnostics that
clearly identify the problem.

Yes there surely are problems with ambiguous errors though it's common to reference the wrong file too, and you would still have to fix it at that level.

There are other problems with that function, too. Even if you give
spice-sdb no reason to inspect included file contents (use -I
--nomunge), it does so anyway. This can make building a subcircuit
library tricky. If you want the automatic file inclusion for subcircuits
to work you must build the library in topological order, because the
included file must exist at netlist generation time even if you're not
including it at that time, and have no dependence on its contents.

John Doty              Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd@xxxxxxxxx

Well that might not be that hard to fix, but one thing at a time. It seems what we want is for the attributes to show up where we can see them eg. have them promoted. This implicitly suggestes that they need values filled in. We also need the option of deleting them without them coming back to haunt us from the inherited attributes.

I think in my original post I suggested something like this.

	Clif


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user