Re: gEDA-user: refdes autonumbering and renumbering

[Charles Lepple <clepple@ghz.cc>]

On Thursday 26 September 2002 06:37 pm, Stefan Petersen wrote:
> Anyhow as I see it Ales is basically right. If you have components named
> C1, C2 and C3 and then restarts gschem the next C-component added should
> be namned 1, ie C1.

I feel like I'm going insane, because when I retested the auto-uref stuff, it 
worked as you described it. We'll chalk that up to user error (sorry about 
that one).

> Is this script still in use? I thought it was retired and replaced by
> some perl/awk/shell script thingie?

dunno about that... I've been using a vi search/replace command to remove 
numbers in preparation for executing the autonumber menu option, and I 
suppose that could be easily turned into a sed command.

> It seems like you use this script in "batch mode" or I'm really lost.
> It's intention is to be used interactivly. I haven't followed this
> discussion too close. I thought noone used it anymore.

to clarify:

I was using auto-uref.scm by adding the following lines in ~/.gEDA/gschemrc :

 (load (string-append gedarcdir "/scheme/auto-uref.scm"))
 (add-hook! add-component-hook auto-uref)
 (add-hook! copy-component-hook auto-uref)

Not "batch mode", as the script was triggered only when I added (or copied) 
components. I hope the "source=subsheet.sch" part in my original post didn't 
mislead you-- all I was saying there was indicating the fact that one symbol 
has a source= attribute, which leads to a subsheet with reference designators 
that are effectively ignored (although the gschem log indicates that the 
subsheet is being read during renumbering operations).

then, I tried the autonumbering menu option (not related to the auto-uref.scm 
code itself, just a similar thing from a (somewhat confused) user's 
perspective).

The other autonumbering problem still exists.

To isolate any oddness caused by the heirarchical design, I narrowed the 
problem down to this testcase:

I opened up a new schematic, placed a whole bunch of resistors (with auto-uref 
disabled, this time), and saved it. All of them had 'refdes=R?'. I then 
executed the autonumber command (menu: "Attributes" -> "Autonumber text...", 
then entered "refdes=R" in the box), and saved it. Everything works as 
expected to this point.

I then close down gschmem. Next, I reopen the schematic, place a few 
resistors, and choose the autonumber menu option again. One resistor will 
have a duplicated refdes attribute (e.g. if R1, R2, R3 and R4 are already 
numbered, then the renumbering process will create a refdes=R4 attribute on 
the first unnumbered resistor (and the rest will have increasing numbers as 
expected).

I've done this test several times with different numbers of resistors labeled 
and unlabeled, and I can send test cases if you want.

The rationale behind doing things this way is as follows: I would like to 
autonumber reference designators, but I frequently quit and restart gschem 
later (hence, auto-uref is not as helpful since it starts over at 1). When I 
execute the autonumber menu command, I don't want to have to remove the 
existing reference designators from components already placed in the 
schematic. (This accounts for the mix of unnumbered and numbered components).

thanks for your patience on this one...

-- 
Charles Lepple <clepple@ghz.cc>
http://www.ghz.cc/charles/