ctrl+d ought result in EOF in ncdirect_readline()

[dankamongmen]

Currently, ctrl+d in ncdirect_readline() just results in a capital D that can be erased and composed. It ought instead trigger an EOF. I'm undecided as to whether any existing text ought be delivered first.

See dankamongmen/growlight#154.

[dankamongmen]

we have:

   if(id == NCKEY_EOF || id == NCKEY_ENTER){                                                                                       
      if(id == NCKEY_ENTER){                                                                                                        
        if(fputc('\n', n->ttyfp) < 0){                                                                                              
          free(str);                                                                                                                
          return NULL;                                                                                                              
        }                                                                                                                           
      }else if(wused == 1){ // NCKEY_EOF 

shouldn't this be catching it?

[dankamongmen]

perhaps relatedly, cli1 < /dev/null just sits there, rather than exiting.

[dankamongmen]

perhaps relatedly, cli1 < /dev/null just sits there, rather than exiting.

and ctrl+cing out of cli1 crashes, jeepers =\

[dankamongmen]

palette_cb:1296:index 145 0xafafaf
process_escape:1877:walk result on 92 (\): 2 428
process_escape:1865:initialized automaton to 1
process_escape:1877:walk result on 93 (]): 0 399
process_escape:1877:walk result on 52 (4): 0 419
process_escape:1877:walk result on 59 (;): 0 420
process_escape:1877:walk result on 49 (1): 0 421
process_escape:1877:walk result on 52 (4): 0 421
process_escape:1877:walk result on 54 (6): 0 421
process_escape:1877:walk result on 59 (;): 0 422
process_escape:1877:walk result on 114 (r): 0 423
process_escape:1877:walk result on 103 (g): 0 424
process_escape:1877:walk result on 98 (b): 0 425
process_escape:1877:walk result on 58 (:): 0 426
process_escape:1877:walk result on 97 (a): 0 427
process_escape:1877:walk result on 102 (f): 0 427
process_escape:1877:walk result on 97 (a): 0 427
process_escape:1877:walk result on 102 (f): 0 427
process_escape:1877:walk result on 47 (/): 0 427
process_escape:1877:walk result on 97 (a): 0 427
process_escape:1877:walk result on 102 (f): 0 427
process_escape:1877:walk result on 97 (a): 0 427
process_escape:1877:walk result on 102 (f): 0 427
process_escape:1877:walk result on 47 (/): 0 427
process_escape:1877:walk result on 100 (d): 0 427
process_escape:1877:walk result on 55 (7): 0 427
process_escape:1877:walk result on 100 (d): 0 427
process_escape:1877:walk result on 55 (7): 0 427
process_escape:1877:walk result on 27 ( ): 0 429
palette_cb:1296:index 146 0xafafd7
process_escape:1877:walk result on 92 (\): 2 428
process_escape:1865:initialized automaton to 1
process_escape:1877:walk result on 93 (]): 0 399
process_escape:1877:walk result on 52 (4): 0 419
block_on_input:2133:blocking on input availability
block_on_input:2137:nonblocking read to check for completion
block_on_input:2178:output queues full; blocking on ipipes
block_on_input:2190:waiting on 2 fds (ibuf: 0/8192)
block_on_input:2217:poll returned 0
block_on_input:2231:got events: ti
growlight-readline: ./src/lib/in.c:1852: process_escape: Assertion `ictx->amata.used < buflen' failed.
;147;rgb:afaf/afaf/ffff^[\^[]4;148;rgb:afaf/d7d7/0000^[

[dankamongmen]

this followed a 2048 byte read

[dankamongmen]

so we need to be setting the input fd persistently high on the EOF case, for one.