Closed JCGoran closed 1 month ago
Attention: Patch coverage is 41.89189%
with 43 lines
in your changes are missing coverage. Please review.
Project coverage is 85.50%. Comparing base (
341b89e
) to head (d224278
). Report is 3 commits behind head on master.
Files | Patch % | Lines |
---|---|---|
src/codegen/codegen_neuron_cpp_visitor.cpp | 20.37% | 43 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Logfiles from GitLab pipeline #208975 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #209020 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #209643 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #210705 (:no_entry:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #211195 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #211286 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #211428 (:white_check_mark:) have been uploaded here!
Status and direct links:
Further comments on: https://github.com/BlueBrain/nmodl-references/pull/20
One should probably check if multiple tables are permitted and work. The reason is that we do use both sig
and sigmoid
to mangle names. Therefore, the question is can one have two different proceedures that both compute (a) sig
(with different values) and both tabulate their sig
.
Logfiles from GitLab pipeline #212706 (:no_entry:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #212725 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #213066 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #213318 (:white_check_mark:) have been uploaded here!
Status and direct links:
One should probably check if multiple tables are permitted and work. The reason is that we do use both
sig
andsigmoid
to mangle names. Therefore, the question is can one have two different proceedures that both compute (a)sig
(with different values) and both tabulate theirsig
.
It seems this will not error out at the parsing stage, but will at the compilation stage:
NEURON {
SUFFIX minimal
RANGE v1, v2
GLOBAL c1, c2
}
PARAMETER {
c1 = 1
c2 = 2
}
ASSIGNED {
v1
v2
}
PROCEDURE p1(arg) {
TABLE v1, v2 DEPEND c1, c2 FROM -1 TO 1 WITH 100
v1 = c1 * arg + c2
v2 = c2 * arg + c1
}
PROCEDURE p2(arg) {
TABLE v1, v2 DEPEND c1, c2 FROM -5 TO 5 WITH 2000
v1 = c1 * arg
v2 = c2 * arg
}
due to redefinitions of v1
and v2
tables:
-> Compiling arm64/ex.cpp
arm64/ex.cpp:220:17: error: redefinition of '_t_v1'
static double *_t_v1;
^
arm64/ex.cpp:218:17: note: previous definition is here
static double *_t_v1;
^
arm64/ex.cpp:221:17: error: redefinition of '_t_v2'
static double *_t_v2;
^
arm64/ex.cpp:219:17: note: previous definition is here
static double *_t_v2;
^
2 errors generated.
so I think it's safe to assume NOCMODL does not support duplicate TABLE variables. I'll see if I can add this check in NMODL as part of one of the visitors. EDIT Fixed in #1280
Logfiles from GitLab pipeline #213586 (:white_check_mark:) have been uploaded here!
Status and direct links:
Logfiles from GitLab pipeline #214372 (:white_check_mark:) have been uploaded here!
Status and direct links:
Needs #1260.