Conversation
…out reference to canfestival
This reverts commit 75682e8.
LMOleksandr-Radchenko
left a comment
There was a problem hiding this comment.
Scary stuff
Also, if the PR is in draft, the jira ticket should not be "in review"
| elif typeinfos.ctype == "visible_string": | ||
| ctx["subIndexType"] = "char" |
There was a problem hiding this comment.
I would not go as far as to say that visible_string is a char. It should either be char* (which probably would cause issues in other places?) or the tool should ignore it with an error (I don't think there are those in the callback right now, are there?
There was a problem hiding this comment.
the global object which are of type_visible string are defined with a char array and not char*
There was a problem hiding this comment.
Here I will just assume you tested it somehow, as I still don't follow how char is the same as char[]
|
We have decided that we would prefer to merge these changes into main if @sveinse and @aleksander-ferkingstad are okay with that. If not we can just keep this branch alive and change our build system to use it. |
|
This is a public tool, so we will not integrate these changes into the python-objdictgen tool. |
I added command line arg that will give the option to generate an extra set of .h/.c files which do not include any reference to canfestival (the #include "data.h" and the types).
I ran odg convert on the same .od file in main and in my branch and a diff shows that there is no difference between before and after my changes. I've attached the files to this PR, notice that the only diff is in the #ifndef sections.
I had to rename the files because GitHub does not allow me to upload .c and .h files
abc_new_c.txt
abc_new_h.txt
abc_new_objectdefines_h.txt
abc_objectdefines_h.txt
abc_c.txt
abc_h.txt