Lambda and getvar
@carlos_m_gil_p wrote:
...You must have vl-catch-all-apply in case the variable does not exist.
Thanks.
Why the need for vl-catch-all-apply? if the variable does not exist then don't include it on the list to process,
This should be enough.
(mapcar '(lambda (x) (cons x (getvar x))) lista)
Unless you meant something different by "subtract", I'm assuming you meant "extract" the current values?
If not, what to subtract the values with?
@pbejse wrote:
...This should be enough.
(mapcar '(lambda (x) (cons x (getvar x))) lista)
....
That was exactly what I came up with, and even if somehow a non-existent System Variable name gets into the lista variable, it doesn't cause any error, but the name simply ends up in a sub-list by itself, with no value paired with it. Applying it to this list:
(setq lista '("whatever" "cmdecho" "osmode" "meshtype" "3dosmode" "orthomode"))
returns this:
(("whatever") ("cmdecho" . 1) ("osmode" . 163) ("meshtype" . 1) ("3dosmode" . 129) ("orthomode" . 0))
If something like that might happen, and you don't want such items included in the result, you can eliminate listed pseudo-variable names that return nil, for example this way [I'm sure there are other ways]:
(vl-remove nil (mapcar '(lambda (x) (if (getvar x) (cons x (getvar x)))) lista))
But where would a non-existent System Variable name come from to be in the list?
@Kent1Cooper wrote:
...it doesn't cause any error, but the name simply ends up in a sub-list by itself, with no value paired with it. Applying it to this list:
True, it does evaluate to nil.
Use foreach instead of mapcar to eliminate the non-existent System Variable name
(setq lista '("cmdecho" "banana" "3dosmode" "osmode" "meshtype" ))
(foreach itm lista
(if (setq f (getvar itm))
(setq nlist (cons
(cons itm f) nlist)))
nlist)
(("meshtype" . 1) ("osmode" . 16495) ("3dosmode" . 139) ("cmdecho" . 1))
@Kent1Cooper wrote:
.... But where would a non-existent System Variable name come from to be in the list?
That is the question.
Hi,
There's a nice routine to do all that stuff
It's called ai_sysvar and it's part of acad20XXdoc.lsp (if I remeber well, it was in acad.mnl in older versions).
It's a nice example of using of defun-q to store the data within the defun-q list.
@pbejse wrote:
@Kent1Cooper wrote:
...it doesn't cause any error, but the name simply ends up in a sub-list by itself, with no value paired with it. Applying it to this list:
True, it does evaluate to nil.
Use foreach instead of mapcar to eliminate the non-existent System Variable name
(setq lista '("cmdecho" "banana" "3dosmode" "osmode" "meshtype" )) (foreach itm lista (if (setq f (getvar itm)) (setq nlist (cons (cons itm f) nlist))) nlist) (("meshtype" . 1) ("osmode" . 16495) ("3dosmode" . 139) ("cmdecho" . 1))
@Kent1Cooper wrote:.... But where would a non-existent System Variable name come from to be in the list?
That is the question.
Be careful by hiding the errors this way.
Probably nobody should write "whatever" or "banana" but anyone can make a typo and write "odmode". In this case, we will not be warned of the typo, the routine can run without error but the "osmode" sysvar will never be reset.
IMO, the user (a developer in this case) have to be warned for such error.
@_gile wrote:
Be careful by hiding the errors this way.
Probably nobody should write "whatever" or "banana" but anyone can make a typo and write "odmode". In this case, we will not be warned of the typo, the routine can run without error but the "osmode" sysvar will never be reset.
IMO, the user (a developer in this case) have to be warned for such error.
In that case, as a developer, The tool "Break on error" & "Last Break Source" in VLIDE is your friend.
Still begs the question of ...where would a non-existent System Variable name come from to be on the list if not from the developer?
EDIT: come to think of it, you are right, it will not show as it will not error., Guess its just good 'ol eagle eyes then 🙂
