Hi all, Stuart Brorson writes: > [...] > I did a quick test of Carlos' stuff. It compiles just fine, and I was able to > run gschem and gnetlist without error. Therefore, I stuck it in CVS. I think it would be even better if we can get ride of the s_scratch_* stuff (from libgeda). There is no real value added by these functions that build an array of strings unique or not. It is also memory hungry and last but not least it is based on a nasty hack. Therefore, I have modified the section where this code is used: - get_selected_component_attributes() in gschem/g_funcs.nw; - g_get_packages() and g_get_non_unique_packages() in gnetlist/g_netlist.c; You will find a diff for each package attached to this message. With these changes it is now possible to delete s_scratch.nw from libgeda. BTW, get_selected_component_attributes() and get_selected_filename() have very complicated prototypes: second and third arguments are never used and first may already be of type TOPLEVEL*. Any particular reason for that (except dependency on now dead code)? Let me know what you think and I will commit these changes to the CVS if everything is ok. And if possible, please test the diffs: I am not really familiar with the internals of gnetlist. Regards, Patrick
Attachment:
gschem.diff
Description: Binary data
Attachment:
gnetlist.diff
Description: Binary data