Re: Exclude pg_largeobject form pg_dump

Guillaume Lelarge <guillaume@lelarge.info>

From: Guillaume Lelarge <guillaume@lelarge.info>
To: amul sul <sulamul@gmail.com>
Cc: Andreas Joseph Krogh <andreas@visena.com>, PostgreSQL General <pgsql-general@postgresql.org>
Date: 2016-11-06T20:33:20Z
Lists: pgsql-hackers, pgsql-general

Attachments

2016-11-04 9:35 GMT+01:00 Guillaume Lelarge <guillaume@lelarge.info>:

> Hi Amul,
>
> 2016-11-04 7:52 GMT+01:00 amul sul <sulamul@gmail.com>:
>
>> Hi Guillaume,
>>
>> I found following issues with this patch, sorry missed in previous post:
>>
>>
> You don't have to be sorry for me doing shitty things :)
>
>
>> #1 :
>>  43 @@ -392,6 +393,10 @@ main(int argc, char **argv)
>>  44                 dopt.outputBlobs = true;
>>  45                 break;
>>  46
>>  47 +           case 'B':           /* Don't dump blobs */
>>  48 +               dopt.include_everything = false;
>>  49 +               break;
>>  50 +
>>
>> Touching dopt.include_everything flag does not seems to be a good idea
>> for '--no-blobs' option, our intension is to exclude blob only, but
>> this excluds other dump too (e.g COMMENT ON DATABASE, CREATE
>> EXTENSION, COMMENT ON EXTENSION, .., etc)  that what we don't want,
>> right?
>>
>>
> Agreed. I was afraid of that, but for some reason, didn't find that. I'll
> fix this.
>

Fixed in v4.


>
>
>> #2 :
>> We should add note for default behaviour if --no-blobs & --blobs both
>> are specified.
>>
>>
> Right. I don't know how I will handle this, but you're right that the
> behaviour should be specified. I'll also fix this.
>
>
I checked other options, such as --format, and there's nothing noted as a
default behaviour. Last one wins, which is what this patch does.


> I'll try to work on this today but as I'm in pgconf.eu 2016, it may be
> only for tomorrow.
>
>
It tooks me more time than expected. v4 attached.

>

-- 
Guillaume.
  http://blog.guillaume.lelarge.info
  http://www.dalibo.com

Commits

  1. Add --no-blobs option to pg_dump