optional renamimg of compilation buffer

[csantosb]

When using hdlmake and helm-make, for example, compilation buffer name is automatically set. Mandatory renaming breaks this functionality.

With this change, using a nil value as fpga-xilinx-vivado-buf would be enough.

[gmlarumbe]

Hi @csantosb ,

Thanks a lot for your contribution.

I tested these changes in my environment and got this error:
Wrong type argument: stringp, nil

The following patch in fpga-utils-define-compile-fn seems to fix it:

diff --git a/fpga-utils.el b/fpga-utils.el
index 9f57ab3..ae15f0a 100644
--- a/fpga-utils.el
+++ b/fpga-utils.el
@@ -166,7 +166,7 @@ COMP-MODE is the name of the compilation derived mode created by macro
   (declare (indent 1) (debug 1))
   `(defun ,name (command)
      ,docstring
-     (when (get-buffer ,buf)
+     (when (and ,buf (get-buffer ,buf))
        (if (y-or-n-p (format "Buffer %s is in use, kill its process and start new compilation?" ,buf))
            (kill-buffer ,buf)
          (user-error "Aborted")))

Could you try it in your environment and add it to the PR if it works well?

Thanks!

[csantosb]

Could you try it in your environment and add it to the PR if it works well?

I push a variant of yours, where I think I stay closer to the original code. I mean: renaming compilation buffers should be optional, but checking if a previous compilation buffer exists is not.

[gmlarumbe]

I push a variant of yours, where I think I stay closer to the original
code.

I do not think you can hardcode *vivado* in the fpga-utils-define-compilation-mode macro since it is also used for other vendors. For example, buf in that macro also refers to fpga-siemens-vsim-buf or to fpga-cadence-xrun-buf.

If you want to use the default compilation buffer names (i.e. *compilation*) with these functions by setting its fpga-<vendor>-<tool>-buf customizable variable to nil (e.g. fpga-xilinx-vivado-buf) while preserving the option of using the default name (i.e. *vivado*) for other users, it is needed to check first if the buffer exists to keep using it for subsequent compilations and to avoid errors with get-buffer since it requires a string as an argument.

However I might be missing something. Let me know if that is the case.

[csantosb]

You're right, your patch applies and fits my needs. To me, this is more than enough.

[gmlarumbe]

You're right, your patch applies and fits my needs. To me, this is more than enough.

Thanks! :)

Do you prefer to add the patch to this PR and rebase the changes or should I merge it and fix it in a subsequent commit?

[csantosb]

Do you prefer to add the patch to this PR and rebase the changes or should I merge it and fix it in a subsequent commit?

Please do ps. thanks, your repos help a lot !

[gmlarumbe]

I'm glad it helps Emacs users :)

Thanks a lot for your contribution!