Formatter: Align multiline string initializer

[mglb]

Demo:

string seq_names[] = {"uart_sanity_vseq",
                      "uart_tx_rx_vseq",
                      "uart_fifo_full_vseq",
                      "uart_fifo_overflow_vseq",
                      "uart_fifo_reset_vseq",
                      "uart_common_vseq",  // for intr_test
                      "uart_intr_vseq",
                      "uart_noise_filter_vseq",
                      "uart_rx_start_bit_filter_vseq",
                      "uart_rx_parity_err_vseq",
                      "uart_tx_ovrd_vseq",
                      "uart_perf_vseq",
                      "uart_loopback_vseq"};

string abc[] = {"a",
                "b",  //
                "c"};

string abc[] = {"a",
                "b",  //
                "c"  //
                };

string abc[] = {  //
  "a",
  "b",
  "c"
};

string seq_names_without_comments[] = {"uart_sanity_vseq",
                                       "uart_tx_rx_vseq",
                                       "uart_fifo_full_vseq",
                                       "uart_fifo_overflow_vseq",
                                       "uart_fifo_reset_vseq",
                                       "uart_common_vseq",
                                       "uart_intr_vseq",
                                       "uart_noise_filter_vseq",
                                       "uart_rx_start_bit_filter_vseq",
                                       "uart_rx_parity_err_vseq",
                                       "uart_tx_ovrd_vseq",
                                       "uart_perf_vseq",
                                       "uart_loopback_vseq"};

string abc_without_comments[] = {"a", "b", "c"};

string seq_names_width_doubled_strings_and_a_variable_name_that_is_too_long[] = {
  "uart_sanity_vseq uart_sanity_vseq",
  "uart_tx_rx_vseq uart_tx_rx_vseq",
  "uart_fifo_full_vseq uart_fifo_full_vseq",
  "uart_fifo_overflow_vseq uart_fifo_overflow_vseq",
  "uart_fifo_reset_vseq uart_fifo_reset_vseq",
  "uart_common_vseq uart_common_vseq",  // for intr_test
  "uart_intr_vseq uart_intr_vseq",
  "uart_noise_filter_vseq uart_noise_filter_vseq",
  "uart_rx_start_bit_filter_vseq uart_rx_start_bit_filter_vseq",
  "uart_rx_parity_err_vseq uart_rx_parity_err_vseq",
  "uart_tx_ovrd_vseq uart_tx_ovrd_vseq",
  "uart_perf_vseq uart_perf_vseq",
  "uart_loopback_vseq uart_loopback_vseq"
};

string aaaabbbbcccc[] = {"aaaaaaaaaaaaaaaaaaaaaaaa",
                         "bbbbbbbbbbbbbbbbbbbbbbbb",  //
                         "cccccccccccccccccccccccc"};

string aaaabbbbcccc[] = {"aaaaaaaaaaaaaaaaaaaaaaaa",
                         "bbbbbbbbbbbbbbbbbbbbbbbb",
                         "cccccccccccccccccccccccc"};

string seq_names_with_non_string_item[] = {
  "uart_sanity_vseq",
  UART_TX_RX_VSEQ,
  "uart_fifo_full_vseq",
  "uart_fifo_overflow_vseq",
  "uart_fifo_reset_vseq",
  "uart_common_vseq",  // for intr_test
  "uart_intr_vseq",
  "uart_noise_filter_vseq",
  "uart_rx_start_bit_filter_vseq",
  "uart_rx_parity_err_vseq",
  "uart_tx_ovrd_vseq",
  "uart_perf_vseq",
  "uart_loopback_vseq"
};

string abc[] = {
  "a",
  B,  // non-string-literal
  "c"
};

string abc[] = {"a", B, "c"};

string nesting[] = {
  "uart_sanity_vseq",
  "uart_tx_rx_vseq",
  {
    "uart_fifo_full_vseq",
    "uart_fifo_overflow_vseq",
    "uart_fifo_reset_vseq",
    "uart_common_vseq",  // for intr_test
    "uart_intr_vseq",
    "uart_noise_filter_vseq",
    "uart_rx_start_bit_filter_vseq",
    "uart_rx_parity_err_vseq",
    "uart_tx_ovrd_vseq"
  },
  "uart_perf_vseq",
  "uart_loopback_vseq"
};

string abcd[] = {
  "a",
  {
    "b",  //
    "c"
  },
  "d"
};

string abcd[] = {"a", {"b", "c"}, "d"};

string seq_names_with_long_non_first_item[] = {
  "uart_sanity_vseq",
  "uart_tx_rx_vseq",
  "uart_fifo_full_vseq",
  "uart_fifo_overflow_vseq",
  "uart_fifo_reset_vseq",
  "uart_common_vseq",
  "uart_intr_vseq",
  "uart_noise_filter_vseq",
  "uart_rx_start_bit_filter_vseq aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
  "uart_rx_parity_err_vseq",
  "uart_tx_ovrd_vseq",
  "uart_perf_vseq",
  "uart_loopback_vseq"
};

Closes #41

[tgorochowik]

@hzeller could you please take a look at this too?

[hzeller]

Nice!

[hzeller]

Below, this value is essentially depending on the result of the !wrap_first_subpartition && (fit_result = FitsOnLine(first_line, style)).fits == true) if condition (if I see it correctly).
So maybe we can assign the value direclty and make it const.

  const bool wrapped_first_subparition = ....

[mglb]

Done

[hzeller]

I wonder if some of these strings would be a bit more readable using c++11 raw string literals to have the expected string shine through more with less escape-backslashes needed.
In general, the strings are fine, leaving it up to you to experiment with that option.

[mglb]

Raw literals could be used only for strings that don't have \n - otherwise we'll need an extra "\n" which are put in separate lines by autoformatter. I think it's better to avoid those few "different" lines and keep everything uniform.
Or did you mean multiline raw strings? Like:

    {R"(string abc[]= {//
"a", "b", "c"
};
)",
     R"(string  abc[] = {  //
  "a",
  "b",
  "c"
};
)"}

[tgorochowik]

I think we can merge this

[hzeller]

This crashes when run over all ibex files; in the following example https://github.com/lowrisc/ibex was checked out to /tmp/ibex/

bazel build verilog/tools/formatter:verible-verilog-format
bazel-bin/verilog/tools/formatter/verible-verilog-format --inplace $(find /tmp/ibex -name "*.sv")
F20210825 14:07:17.748425 152869 tree_utils.h:75] Check failed: E(node.Tag().tag) == expected_node_enum (kClassNew vs. kExpression) 
*** Check failure stack trace: ***
    @     0x55e080ced280  google::LogMessage::Fail()
    @     0x55e080ced1cb  google::LogMessage::SendToLog()
    @     0x55e080cecaf2  google::LogMessage::Flush()
    @     0x55e080cf02d8  google::LogMessageFatal::~LogMessageFatal()
    @     0x55e080bf1658  verible::CheckNodeEnum<>()
[...]

Bisecting shows that this is the commit where that happens. Can you roll back @tgorochowik @mglb ? (since there were other commits afterwards, it is not a simple reverse patch unfortunately).

[hzeller]

(or fixing and roll-forward is also an option of course)

[tgorochowik]

Sure, we'll take a look at this