Fix https://github.com/ra3xdh/qucs_s/issues/967:

[ThomasZecha]

-The lookahead-pattern for differentiate the 3-/4-/5-node spice-bjt
erroneously contains newline. This causes read a spice bjt-line don't
stop at new line and detect a 5-node bjt instead of 3-node bjt as
happen for AD822X.cir of Opamp_AC_Tran.zip from the issue 967.
This is fixed and tested against AD822X.cir. The critical lines from
AD822X.cir are included in the bjt.cir qucsconv_rf testing example.

cherry-picked from 3c375fd:

Fix build issues
-fixed sprintf format buffer overflow warning: check_citi.cpp,
check_mdl.cpp, check_spice.cpp
-fixed signed compare/expression warning: check_mdl.cpp, circuit.cpp,
hbsolver.cpp, e_trsolver.cpp, nasolver.cpp, spline.cpp, tmatrix.cpp,
tvector.cpp
-fixed clearing of non-trivial type via memset warning by static_cast:
circuit.cpp, tmatrix.cpp
-fixed int-in-bool-context warning: thyristor.cpp
-fixed unnecessary parentheses warning: environment.cpp, equation.cpp,
evaluate.cpp
-fixed case fallthrough warning: equation.cpp
-fixed unused parameter/variable warning: e_trsolver.cpp, real.cpp, qucs_action.cpp
-fixed infinite recursion warning: real.cpp, real.h
-fixed format type warning: module.cpp
-fixed depracated copy warning by adding assignment operator: strlist.cpp,
strlist.h
-fixed misleading indentation warning: vector.cpp
-fixed deprecated yacc/bison name-prefix directive warning: parse_spice.ypp,
parse_vcd.ypp, parse_citi.ypp, parse_csv.ypp, parse_dataset.ypp, parse_mdl.ypp
parse_mdl.ypp parse_netlist.ypp, parse_touchstone.ypp parse_zvr.ypp
-fixed useless yacc/bison symbol definition: parse_vcd.ypp
-fixed shift/reduce-conflicts (sr) and reduce/reduce-conflicts (rr) of the spice-parser
build by parse_spice.ypp and scan_spice.lpp:
-sr-conflicts solved by introducing token/rule precedence
-rr-conflicts solved by introducing token lookahead's in the spice scanner
-verification of equalness of generated output from both the converter before and
after the code change done with spice input bjt.cir done
-fixed shift/reduce-conflicts (sr) and reduce/reduce-conflicts (rr) of the csv-parser
build by parse_csv.ypp and scan_csv.lpp:
-sr/rr-conflicts solved by downsize and reorder grammatic rules
-verification of equalness of generated output from both the converter before and
after the code change done with csv input qucs_csv.csv done
-removed useless scanner token definition and replaced misleading operator assocativity
with precedence: parse_netlist.ypp
-fixed shift/reduce-conflicts (sr) of the touchstone-parser build by parse_touchstone.ypp and
scan_touchstonelpp:
-sr conflicts solved by remove useless/redundant grammatic rules
-verification of equalness of generated output from both the converter before and
after the code change done with touchstone input R50C1pF.s1p done

[ThomasZecha]

@ra3xdh: Hi Vadim, it seems that the PR will not be accepted. May the the change set is to big, or the error risk is not acceptable? Is there something I can do?
From my point of view all the build issues shall be fixed soon or later, and this PR is a big step to reach it. I would go even further and switch to Werror after a warning-free build status.
Thanks.

[ra3xdh]

This PR may be accepted. Please split this patch at least into two parts:

• The fixes related to qucsator simulator itself (C++ build warnings)
• The fixes related to qucsconv flex/bison parsers

The qucsconv is not checked by automatic tests and it requires to test every combination by hands. See also #27. I will try to provide a review in the next weeks. Don't expect a quick response, because I am short in time in this month.

[ThomasZecha]

Change-set splitted now in 1 c++-part commit covering all build-issue fixes for c++ source code and 7 flex/bison part commits dedicated to the different changed parsers and general topics for fix flex/bison build issues.

[ra3xdh]

I have seen the spitted PR. I will provide a review within the next month.

[ThomasZecha]

Any blocking points?

[ra3xdh]

No blocking points. The qucsconv related part requires a manual testing. I am time constrained now and try to perform this task in January. You may speed up the merging if you provide a patch implementing an automatic test to qucs-test suite. See #27