Make use of variables

garyd9

In regards to the proposal...

There are currently several different "standards" for gcodes, but they are (for the most part) compatible with each other. I can usually take a basic gcode file that prints a cube and run it on any printer's firmware (or firmware specific preprocessor such as the case of sailfish) that supports gcode. 95% of the time, that gcode file will run fine. Even if a particular flavor of gcode doesn't support a specific command, it will usually recognize that it's not supported, spit out a warning, and continue.

This proposal isn't a case of adding some M commands that are specific to RRF (but could be safely ignored by other processors.) This proposal is a fundamental change in the gcode "language" while trying to still call it gcode. For example, I believe that gcode defines the "G1" command as one single linear movement (potentially in more than one axis.) With this proposal, however, that G1 command might be inside a loop (that only RRF understands) and really means to make one movement 2 times, 10 times, or a hundred times.

To add to the issue, there's no gcode command (that I know of) that could tell a gcode processor "don't run this file as it can't be compatible with you." I suppose you could have a header that includes a "M0" command wrapped with an always false conditional similar to:

if 1 = 0
M0
end if

If a gcode processor ignores any line it doesn't understand, it would only process the "M0" and stop. (Assuming it supports M0 to begin with. Not all gcode processors do.)

Anyway, to get to the point, I'd suggest NOT making changes to gcode that actually change the language. If you want to write a new language, then write a new language. Not only would this prevent compatibility issues (by assuring incompatibility) but would also free you from any constraints of trying to be gcode compatible when it clearly can't be.

dc42

I anticipate that the new features would be used almost exclusively in macro files that are used only on printers running RepRapFirmware.

garyd9

@dc42 said in Make use of variables:

I anticipate that the new features would be used almost exclusively in macro files that are used only on printers running RepRapFirmware.

Then perhaps call it something else (RRF-Code or Duet-Code) with a different file extension and only allow it in macros. I realize that you have a certain expectation of competence of your users, but there's a large potential for something to go wrong when these machines are being used for laser cutters, CNC, etc

Googliola

@dc42 From a user perspective, the terminology would cause confusion and additional and maintained documentation , but then the question is how much you weigh user-friendliness.
If I were in your shoes, I would probably stick to ease and coherence with known / documented guidelines. You might open Pandora box here. People might argue that common (Fanuc) standards are 'better'. Comes down to taste at the end of the day really...

bot

I really like your proposal, dc42.

I have only one immediate use-case that I would want to employ: check current z height, and raise then lower bed to switch IDEX tools and avoid obstacles near the bed perimeter, if current z height is insufficient to clear said obstacles. This would have to be located in the actual print gcode files, in the tool switching script.

I don't see any problem with doing this, based on your proposal, but just wanted to mention a use-case that would have the script inside a gcode file for printing.

edit: Upon further reflection, these commands could be contained in the firmware-based macro tool change scripts. I was thinking of a case of using slicer based tool change scripts, but that would not be necessary. Ignore my stated use case of scripts in a printable gcode file.

Googliola

@dc42 Haven't had much time lately to follow up on my feature request. With holidays ahead, this will change soon

Just been wondering how the object model is coming along and if you could make a statement about when / what version it will be implemented.
Thanks in advance

dc42

@googliola said in Make use of variables:

@dc42 Haven't had much time lately to follow up on my feature request. With holidays ahead, this will change soon

Just been wondering how the object model is coming along and if you could make a statement about when / what version it will be implemented.
Thanks in advance

A small amount of the object model is supported in firmware 2.02RC5 along with variable substitution, as a proof of concept. The full implementation is planned for release 2.03.

fma

@garyd9 said in Make use of variables:

@dc42 said in Make use of variables:

I anticipate that the new features would be used almost exclusively in macro files that are used only on printers running RepRapFirmware.

Then perhaps call it something else (RRF-Code or Duet-Code) with a different file extension and only allow it in macros. I realize that you have a certain expectation of competence of your users, but there's a large potential for something to go wrong when these machines are being used for laser cutters, CNC, etc

Didn't follow this thread before... I agree with @garyd9: as no slicer will generate such custom G-Code, if the goal is to use it from macros, I suggest to write a macro language, like Slic3r does, evaluated by the macro pre-processor, and keep a standard G-Code.

dc42

The additional facilities will be provided by way of additional (not replacement) syntax and will not affect the ability of RRF to run GCode produced by slicers.

fma

Yes, I understand, but this means that it will be possible to use this syntax from Slicers (using custom code), which is, IMO, more dangerous than from macros only...

fma

Maybe another thing to think about is having the possibility to give params to macros...

dc42

@fma said in Make use of variables:

Maybe another thing to think about is having the possibility to give params to macros...

Yes, I will probably include that.

Ringo1508

Glad I found this topic! Sorry for dragging up an old topic but I don't see any other information on the wiki about this.

I'm looking for variable support as well. An example of an immediate want (notice I didn't say need lol) is the ability to call a macro with hotend first layer temperature and bed temperature parameters.

My Cartesian machines all will set initial hotend temp to 160° while bed heats up, then do a mesh bed leveling and finally heating up the nozzle to print temp just outside the print area and doing a purge line before it begins the print.

Yes I know that I can put this in Slic3r custom gcode for print start, but then I have to manage it for each printer I'd love to take advantage of just putting the macro call in the custom gcode section.