Re: gEDA-user: text labels on symbols, DRY, colon cancer, and redundant, repetitive redundancy.
Would anyone be opposed to having the netlister assume a :1 for net= attributes without a colon? There was a patch for this available, but it did not make it in the official code. I would love it if I didn't need the :1 on netlist names. Right now I do not have :1 on my net names, and I have my makefile make a copy of all my schematics, and run sed -i -e 's/^\(net=.*\)/\1:1/' on them to add a :1. It works okay, but it's annoying that there are duplicate files created, and some makefile rules (creating a pdf or png) need to use the original files, while others (DRC, making a netlist) need to use the modified files. - Miles On Thu, Jun 11, 2009 at 1:18 PM, Stefan Salewski m...@ssalewski.de wrote: On Thu, 2009-06-11 at 13:47 -0500, Mark Rages wrote: In http://geda.seul.org/wiki/geda:scg, referring to the device= attribute: This attribute should not be used as a label. Use a separate text object for the label. What is the reasoning behind this? This is pointless redundancy. And it makes using light symbols harder, because you need different symbols for different device= attributes, or else there is no schematic-visible way to differentiate between them. True. I have replaced the plain text like 74HC04 in my symbols with attributes. I used value attribute. I did this for the reasons you mentioned above. I ask about it on this list -- no reply, so it may be OK. Also, I notice that the power supply symbols have a separate text label and hidden net= label with a :1 appended. In fact, it is necesary to append a :1 to the end of any net= attribute. This is ugly. And it makes the redundant text label necessary to avoid adding the information-free :1 onto the schematic. I asked about this on the list too. Redundancy is indeed bad -- my decision is to make the net= a visible attribute. So the :1 for pin 1 is visible too -- not too ugly, but it may confuse other people. Would anyone be opposed to having the netlister assume a :1 for net= attributes without a colon? There was a patch for this available, but it did not make it in the official code. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Breadboard drawings with pcb?
I have this same problem, because I automate my build with makefiles, and I'd like to have a build machine (that doesn't run X) be able to build everything (mostly software, but also PNGs and PDFs of my schematics, along with running DRC, making a netlist, etc.) I did a search for an alternate way around the problem, and found this: http://stackoverflow.com/questions/834723/a-dev-null-equivilent-for-display-when-the-display-is-just-noise I haven't tried it yet, but plan to soon. - Miles On Tue, May 26, 2009 at 9:35 AM, John Doty j...@noqsi.com wrote: On May 26, 2009, at 5:34 AM, Ineiev wrote: On 5/26/09, Josef Wolf j...@raven.inka.de wrote: But scriptability is a concern, though: is it possible to create the ps/eps from a script/Makefile without GUI intervention? (Do you mean you do it _with_ GUI?) pcb -x ps --psfile board.ps board.pcb PCB even does not requires X for this task. BTW I couldn't achieve this with gschem --- it doesn't work from text terminal for me. You can print using the print.scm from a text terminal or script as long as there's an X server for gschem to flash the page on. A minor annoyance, I think. Regards, Ineiev ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user John Doty Noqsi Aerospace, Ltd. http://www.noqsi.com/ j...@noqsi.com ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
gEDA-user: Xvfb allows 'printing' with gschem with a headless server (was: Breadboard drawings with pcb?)
http://en.wikipedia.org/wiki/Xvfb In one terminal: Xvfb :3 In another terminal: DISPLAY=:3 gschem -p -o myschematic-page1.ps -s print.scm myschematic-page1.sch Xvfb doesn't need any particular hardware resources (screen, video card, keyboard, mouse), and it allowed the above actions to complete without the dreaded Gtk-WARNING **: cannot open display. In my case I am trying to do the above on a machine I don't have root access on, where there's already a display server running on :0, so I just picked a random other display to run Xvfb on. - Miles On Tue, May 26, 2009 at 2:12 PM, Miles Gazic miles.ga...@gmail.com wrote: I have this same problem, because I automate my build with makefiles, and I'd like to have a build machine (that doesn't run X) be able to build everything (mostly software, but also PNGs and PDFs of my schematics, along with running DRC, making a netlist, etc.) I did a search for an alternate way around the problem, and found this: http://stackoverflow.com/questions/834723/a-dev-null-equivilent-for-display-when-the-display-is-just-noise I haven't tried it yet, but plan to soon. - Miles On Tue, May 26, 2009 at 9:35 AM, John Doty j...@noqsi.com wrote: On May 26, 2009, at 5:34 AM, Ineiev wrote: On 5/26/09, Josef Wolf j...@raven.inka.de wrote: But scriptability is a concern, though: is it possible to create the ps/eps from a script/Makefile without GUI intervention? (Do you mean you do it _with_ GUI?) pcb -x ps --psfile board.ps board.pcb PCB even does not requires X for this task. BTW I couldn't achieve this with gschem --- it doesn't work from text terminal for me. You can print using the print.scm from a text terminal or script as long as there's an X server for gschem to flash the page on. A minor annoyance, I think. Regards, Ineiev ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: building gEDA for cygwin
I have 1.4.1.20080929, that's the only version I've tried to build. - Miles On Wed, May 20, 2009 at 7:26 AM, Alberto Maccioni alberto.macci...@gmail.com wrote: Has anyone been able to compile gEDA v1.4.3 with cygwin? I can only build 1.4.2, but not 1.4.1 What about v1.5.x? It needs GTK 2.8 but unfortynately cygwin only has 2.6 ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: fritzing
I agree it's a waste of time to port gEDA to windows, and that a cross-platform library is the way to go. I've used SDL a bit, but I wouldn't pick it for this. I'd (personally) pick Qt, and recommend porting gEDA to Qt, so there'd be one unified code base, instead of a separate codebase for unix+X11 and for (no matter how much code they have in common). I've done Qt apps on windows/Linux/OS X, and it works seamlessly on all 3 platforms. I prefer it to Win32 or MFC apps on windows, and it's also my favorite way to write linux apps, even disregarding cross-platform requirements for either platform. I tried making cross-platform gtk apps, and basically gave up, and instead cross-compiled windows apps from a linux box (something that worked okay for me, but doesn't help people who are windows-only to develop the apps). If you try to port gEDA to windows, you basically will do all the work of creating a cross-platform toolkit, and have to maintain it for ever: why not use an existing one? I'm not saying you should port to Qt or asking someone to do so, or complaining that gEDA doesn't use Qt now. Just offering advice on what to use, if your goal was to make a GUI app run seamlessly on linux/OS X/windows. - Miles PS: I use gEDA on cygwin. I have a cygwin mirror at work, and I made a cygwin package for gEDA, as well as making a modified cygwin installer so that coworkers can run the installer to automatically install all the packages I think they need from our local mirror, including my gEDA package. It was a PITA to set up, and I doubt many people at companies even an order of magnitude bigger than where I work will devote the resources to do something like that. And even still after all my work, I couldn't get my package to conform to cygwin's package acceptance criteria, so I never tried to push it upstream. Part of the reason was because I couldn't quite figure out how to get gEDA to build properly, and I had to do ugly hacks to make it run. Someone more familiar with gEDA and/or cygwin than I am would have to make the package for it to get accepted upstream, or I'd have to spend more time figuring stuff out, and I (sadly) haven't gotten around to it. On Sun, May 10, 2009 at 2:50 AM, Michael Torp Kaalund kaal...@gmail.com wrote: If you are going to port a program between operation systems, then it is better to use something like libsdl, that can be used to crosscompile. Now I hav'nt seen the code for the programs in gEDA, but I think it will a while to port the existing code to use something like libsdl, and it is something that definally needs a plan to do. But if you really like gEDA on a windows mashine, then you can try use Cygwin. Thats my 2 cents. Hav a nice saturday.. -Michael Kaalund (Denmark) lør, 09 05 2009 kl. 23:06 -0700, skrev Dave N6NZ: Dave McGuire wrote: Apparently, programming for Windows is much more difficult than programming for unix. I've never seen it, but I'm told that the Windows API is actually rather nice. I suspect it's more a matter of more highly-skilled programmers not working with Windows. At least for the most part. Windows is just way different. There is no where near a 1:1 mapping of functionality in API calls between Windows and *nix. And things that are efficient on one system can be awful on the other... fork() being an example of something that is notoriously expensive on Windows and highly tuned (usually) on *nix. If you find one of those deeply embedded in the architecture of a program, it can make porting very painful. And the Windows API specification is much more volatile. Porting isn't easy. And even if you decide up front to design for portability among Windows, OS X, and X-windows, there is no single, obvious, straight-forward strategy. -dave And don't ask me for more details I've purposely avoided Windows programming (because I could) and have fallen woefully behind on Windows programming methodologies. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: fritzing
For software to be truly expert friendly, it must use languages that are meaningful in the application domain, and lots of extendability. To a circuit designer, that is not C, Scheme, M4, or XML. The ones I know circuit designers use are verilog, perl and python. and then there are the many Matlab programmers... Are any of those what you're thinking of Al? Verilog or VHDL are the only languages that I know of that are directly used in hardware design, and IMO neither is suited for writing scripts on PCs. Since no suitable languages are commonly used in hardware development (except by people who also write software), I don't think there's a clear choice. Any scripting language would be fine, but something like perl or python would be more accessible to more people than scheme or M4 currently is. The main issue though is to have input and output files in plain text, so anyone can write scripts in whatever language they please to process them. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Netlisting a BOM
I think the file you are missing is attribs http://www.geda.seul.org/wiki/geda:faq-gnetlist mine: $ cat attribs device value manufacturer manufacturer_part_number footprint On Thu, Apr 30, 2009 at 1:59 PM, Rob Butts r.but...@gmail.com wrote: I'm trying to use gnetlist -g bom -o... to generate a bill of material. The documentation calls for a config file to tell the netlister what attributes to list but it doesn't say how to name it or use it. Thanks, Rob ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: gEDA attributes -- why not in this simple way
On Thu, Apr 9, 2009 at 9:40 AM, John Doty j...@noqsi.com wrote: Where is the problem by this simple approach? Cognitive dissonance in the minds of users and developers, whose minds actually work hierarchically, but who expect, based on traditional practices, that component management is flat. I guess I'm in the minority camp. Stefan's description made immediate sense to me, but I simply cannot understand the way gschem handles this right now, it seems like a convoluted selfcontradictory mish-mash. Can anyone point me to a doc that explains the 'heirarchical' point of view? I think if I could get a conception of why it works the way it does, I'd probably be able to adapt myself to that kind of workflow. Thanks, Miles ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Why are visible attributes in a .sym not visible in my .sch?
I have attributes in my symbol files that I mark as visible. This makes them visible when editing the .sym file, but NOT when editing the .sch file. I don't think so. I think if I add visible attributes to a symbol, they are visible (and edible) in the schematic. I think I read that visible attributes are always promoted? I don't want the attributes promoted, that would result in multiple copies of the same information that need to be kept synchronized. The attributes that aren't promoted that reside in my .sym files work properly when I create a BOM with gnetlist. The problem is that they aren't visible on my schematic in gschem. It seems to me that promotion is a very different thing than making sym attributes visible in gschem. Promotion makes a *copy* of the attributes when you insert a symbol into the schematic, and puts that in the .sch files. Gschem then sees that copy, which from it's point of view while editing, is just an attribute that exists in the .sch, and has nothing whatsoever to do with the .sym file. If there's something bad about having attributes in a .sym file be visible in a schematic (without copying, aka 'promotion'), then why does gnetlist have access to those attributes the same as attributes that are in the .sch? I'm not trying to be a jackass, I get the feeling there is a reason the creators of gschem did this, and I'd like to know what it is so I can use the tools properly. Maybe you should give an example, and tell use your gschem version. gEDA/gschem version 1.4.1.20080929 What is the benefit of sending the the text of themessage twice? Sorry, that wasn't intentional. Perhaps that's what happens when the list receives an html email. Thanks, Miles ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Why are visible attributes in a .sym not visible in my .sch?
John wrote: Huh? Not the way it works for me (currently using gEDA/gschem version 1.4.3.20081231). Unpromoted visible attributes from the sym are visible in the schematic, you just can't edit them. Are you sure you've marked them visible? Yes, absolutely sure. This happens on 1.4.1.20080929 on WinXP, as well as 1.4.0.20080127 on Ubuntu 8.10. Stefan wrote: I have currently no idea how you manage to have attributes marked visible in symbols, which are invisible in schematic. That was why I ask for an example. Sorry for not including an example earlier. I tried to make a small example to demonstrate the problem. 3 files are attached: gafrc: needed at a minimum to have (component-library .), so the .sym is found IL717-3E.sym: A pared down symbol to demonstrate the problem test.sch: a schematic that has two IL717-3E symbols in it. One was inserted with attribute promotion on, and the manufacturer_part_number attribute was copied to the schematic file (bad, because I don't want duplication of information). The other was inserted with attribute promotion off, and it has no attributes in the .sch (not even an empty {} section). When I open it in gschem, I do not see the text IL717-3E, but I would like to see that text. Thanks very much, Miles gafrc Description: Binary data IL717-3E.sym Description: Binary data test.sch Description: application/schdpl32 ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
gEDA-user: Why are visible attributes in a .sym not visible in my .sch?
I have attributes in my symbol files that I mark as visible. This makes them visible when editing the .sym file, but NOT when editing the .sch file. I do not want them promoted (which I think will copy the text of them to the contents of my .sch files, when a symbol is inserted in my schematic). I *do* however want them to be visible in my schematic. Otherwise, I wouldn't have made them visible in the first place. What I'm confused about is that gschem lets me make attributes visible in my .sym, but that doesn't result in me being able to see them in my schematic. What possible reason is there for marking an attribute visible, if not because I want it to be visible on the schematic? This is too basic a feature to be a bug... Could someone please explain the reason why gschem works this way, and if possible, what I can do to make it work the way I want it to? Bonus points if you explain why the way I want it to work isn't the default, and what's wrong with it. Thanks, Miles ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user