Re: Exportability of EDA industry from North America?
From: Stuart Brorson (sdb_at_cloud9.net)
Date: 01/18/05
- Next message: Rich Grise: "Re: Exportability of EDA industry from North America?"
- Previous message: DG: "Josephson Junction spice models"
- In reply to: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Next in thread: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Reply: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Messages sorted by: [ date ] [ thread ]
Date: Tue, 18 Jan 2005 23:38:04 -0000
In sci.electronics.cad Chuck Harris <cf-NO-SPAM-harris@erols.com> wrote:
: Hi Stuart,
: Ok, I have figured out what's going on. If you cd to the directory
: that the RF_Amp is in, and then invoke gschem, Open page and select
: the MSA-2643.sch drawing, everything works the way you all probably
: expected it to.
: However, if you are in some miscellaneous directory, and invoke gschem,
: Open page, and wind your way through the directory structure until
: you find the MSA-2643.sch drawing, and open it, then you get a schematic
: with blank spots where the transistors should be.
: So, what is happening is you only read the gafrc file for the directory
: that you are in when gschem is invoked.
: What to do? It doesn't seem too unreasonable for a person to be able to
: start up the geda tools from any old directory they happen to be in and
: browse their way to the design files using open page, and expect the
: right thing to happen.
: It seems to me that a reasonable solution would be to have gEDA tools
: read any appropriate rc files found in the directories where you open
: a new, or old page. I am unsure what would be the best solution for
: opening a new page in a new empty directory. Perhaps if the directory
: is created from within the gEDA tools, the directory should be prefilled
: with an rc file from some template directory?
Hi Chuck --
OK, this is a user education issue. Sorry!
Loosely speaking, gEDA has three sets of RC files: One system wide,
one for you as a user, and one living in the local project directory:
* The system one lives in ${PREFIX}/share/gEDA, where ${PREFIX} =
wherever you stuck your gEDA stuff when you installed the dist. This
one holds most of the path info about where symbols live. It is
always read in upon program start-up.
* The user one should live in ${HOME}. You can use this for
customizations you would like to use across all your projects. (Like
remapping the key actions, or screen colors.) Gschem
tries to read it upon program start-up, but only warns you if it can't
find it.
* The project one lives in the project directory where the schematic
files are stored. This one is used for customizations local to that
particular project. (Like storing local symbol files, or other
project-specific stuff.) Gschem tries to read it upon program
start-up, but only warns you if it can't find it.
When gschem opens up, it tries to read all three. However, if you
run gschem on a file living in another directory from the one you are
in, then it doesn't see the project-specific RC file.
In the case of the RF_Amp example, the gafrc file lives in the RF_Amp
project directory. I am not sure that making that project-local RC
file work when reading the .sch file from outside that directory is a
good idea. The vision is that you put project-specific customizations
into there. That means that you run gschem while working out of that
directory.
There is a mechanism for specifying which RC file to load using the -r
flag, but it doesn't seem to follow file paths like "./sym" correctly
(I just checked). I can file a bug report on this.
I agree that the nice thing to do is to have all schematic open
correctly from everywhere. The way to do that is to have gschem
deduce the project directory from the path to the file being opened,
and then open the corresponding RC file. But what if a user then
opened two files in two different directories? Which RC file to open?
Both? What if they conflict?
Therefore, this question needs some further thinking. Meanwhile, it
does raise the question of bullet-proofing the examples for newbies.
I'll think about that.
Stuart
- Next message: Rich Grise: "Re: Exportability of EDA industry from North America?"
- Previous message: DG: "Josephson Junction spice models"
- In reply to: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Next in thread: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Reply: Chuck Harris: "Re: Exportability of EDA industry from North America?"
- Messages sorted by: [ date ] [ thread ]