>> This might interfere with the LZTEXT data type as is. Since
>> this new data type isn't released yet, I'll have to remove it
>> again before freeze. The compressor/decompressor will become
>> part of the toaster.
>>
>> Needs to revert the changes to pg_rewrite too, so the next
>> release will NOT gain any bigger rules. But to release it and
>> later fight problems that I already see, only to enable
>> bigger rules right now, is IMHO the wrong decision.
> Agreed.
Yes. I thought that LZTEXT was redundant as soon as we agreed that
compression could be part of the out-of-line storage mechanism.
Better to pull it now than have to support it long after it has
no purpose. We can struggle along with limited-length rules for
a little bit longer.
regards, tom lane