Greg Stark <stark@mit.edu> writes:
> * Simplify find_my_exec by using realpath(3)
> - Tom Lane is author but I don't know if he intends to apply this in
> this release
I'd like to get it done. It's currently stuck because Peter asked
for some behavioral changes that I was dubious about, and we're trying
to come to a mutually-agreeable idea. However ... the only relation
between that issue and what I actually want to do is that it's touching
the same code. Maybe we should put the behavioral-change ideas on the
shelf and just get the realpath() reimplementation done for now.
I'm getting antsy about letting that wait much longer, because it's
entirely possible that there will be exciting portability problems;
waiting till almost feature freeze to find out doesn't seem wise.
>> * Code checks for App Devs, using new options for transaction behavior
> I think features supporting a "dry-run" mode would be great but
> judging by the lack of response this doesn't look like the version of
> that people are interested in.
This version of it seems quite unsafe. I wonder if we could have some
session-level property that says "no changes made by this session will
actually get committed, but it will look to the session as if they did"
(thus fixing the inability to test DDL that you noted). I'm not sure
how well that could work, or what leakiness there might be in the
abstraction. There would probably be locking oddities at the least.
I agree with RwF, or maybe Withdrawn given that Simon's retired?
The same for his other patches --- somebody else will need to push
them forward if anything's to come of them.
regards, tom lane