Re: pg_dump option to dump only functions

Thanks! Any reason the comment shouldn't also mention this directly? I.E.:
/*
This covers the long options equivalent to -X xxx.
The -X options return 0 if called in long form, because then we store a flag 
rather than returning a distinct value.
See man 3 getopt.
*/

Sean

----- Original Message ----- 
From: "Andrew Dunstan" <andrew@dunslane.net>
To: "Sean Utt" <sean@strateja.com>
Cc: <pgsql-hackers@postgresql.org>
Sent: Sunday, October 30, 2005 3:05 PM
Subject: Re: [HACKERS] pg_dump option to dump only functions


>
> You have omitted the "case 0" line following the comment, which is in fact 
> what it refers to. The -X options return 0 if called in long form, because 
> then we store a flag rather than returning a distinct value. See man 3 
> getopt.
>
> cheers
>
> andrew
>
> Sean Utt wrote:
>
>> In what might be called my spare time, I was looking at pg_dump.c to see 
>> about adding an option to dump only functions, and I think a comment got 
>> pushed out of place in the section for handling arguments:
>>
>>    395                         case 'X':
>>    396                                 if (strcmp(optarg, 
>> "disable-dollar-quoting") == 0)
>>    397                                         disable_dollar_quoting = 
>> 1;
>>    398                                 else if (strcmp(optarg, 
>> "disable-triggers") == 0)
>>    399                                         disable_triggers = 1;
>>    400                                 else if (strcmp(optarg, 
>> "use-set-session-authorization") == 0)
>>    401                                         use_setsessauth = 1;
>>    402                                 else
>>    403                                 {
>>    404                                         fprintf(stderr,
>>    405                                                         _("%s: 
>> invalid -X option -- %s\n"),
>>    406                                                         progname, 
>> optarg);
>>    407                                         fprintf(stderr, _("Try 
>> \"%s --help\" for more information.\n"), progname        );
>>    408                                         exit(1);
>>    409                                 }
>>    410                                 break;
>>    411
>>    412                         case 'Z':                       /* 
>> Compression Level */
>>    413                                 compressLevel = atoi(optarg);
>>    414                                 break;
>>    415                                 /* This covers the long options 
>> equivalent to -X xxx. */
>>                                          ^^^^^^^^^^^^^^^^ --------------  
>> This comment seems out of place here. I imagine it once was after the 
>> break for case: 'X': (line411) and got misplaced when case 'Z': was 
>> added. Any other fantasies about how it got here, or where it belongs?
>>
>> My other fantasy is that it was supposed to go here:
>>    241                 /*
>>    242                  * the following options don't have an equivalent 
>> short option letter,
>>    243                  * but are available as '-X long-name'
>>    244                  */
>>    245                 {"disable-dollar-quoting", no_argument, 
>> &disable_dollar_quoting, 1},
>>    246                 {"disable-triggers", no_argument, 
>> &disable_triggers, 1},
>>    247                 {"use-set-session-authorization", no_argument, 
>> &use_setsessauth, 1},
>> on line 248.....
>>
>> I am not sure where it should go, but it seems pretty out of place where 
>> it is.
>>
>> Sean
>>
>>
>>
>>
>> ---------------------------(end of broadcast)---------------------------
>> TIP 2: Don't 'kill -9' the postmaster
>>
>
>