Unbreak $rs->create() with empty hashref on Oracle

[lasse-unity3d]

Oracle does not support the INSERT INTO ... DEFAULT VALUES syntax used by
SQLMaker by default. Furthermore, Oracle has no way of inserting a row
without specifying any columns, so the only reasonably fix seems to be to
pick a column and supply the DEFAULT keyword as value for it.

This approach seems to work without problems, even for things like
sequence+trigger-as-auto-increment.

[ilmari]

Don't modify the caller's hashref, just do $to_insert = { $col => \'DEFAULT' }

[lasse-unity3d]

Oops. Will fix.

[lasse-unity3d]

2nd attempt - fixed accidental stepping on caller's hashref in insert() as pointed out by @ilmari .

[ilmari]

Merged manually to blead as 236b59c

[ribasushi]

Hi @lasse-unity3d

I had a chance to review your contribution after it was merged to the dev branch, and found some issues. Please read the detailed explanation here: f8135512c

It would be great if you could spend some time to get the thing into shape, however given it has been a while it may be the case you no longer have the opportunity to work on this. Please let us know if this is the case, so someone else can pick it up and slog it to completion.

In any case: thanks for helping us make DBIC better ;)

[lasse-unity3d]

On Wed, Oct 22, 2014 at 9:57 PM, Peter Rabbitson notifications@github.com
wrote:

Hi @lasse-unity3d https://github.com/lasse-unity3d

I had a chance to review your contribution after it was merged to the dev
branch, and found some issues. Please read the detailed explanation here:
f813551 f8135512c

Thanks for the explanation. I'm not nearly familiar enough with DBIC
internals yet to really grasp the implications of what you explained but
I'll try to read up on it.

Do you have concrete examples of breakage resulting from the patch? That
would be really helpful for me to understand the problem better...

It would be great if you could spend some time to get the thing into
shape, however given it has been a while it may be the case you no longer
have the opportunity to work on this. Please let us know if this is the
case, so someone else can pick it up and slog it to completion.

The issue is still relevant to me but since our porting to Oracle has been
postponed it's not my focus at the moment. We're in a bit of a crunch
currently but I'll try to get around to this again soon...

Vacation next week may or may not help with this... :-)

/L

In any case: thanks for helping us make DBIC better ;)

—
Reply to this email directly or view it on GitHub
#61 (comment).

[ribasushi]

This patch uncovered extra irregularities in the base insert() codepth. I am still sifting through those, so the above has been put on hold.

I just wanted to update you that this PR is in my todo list, and will be merged in some manner before the next stable release (likely some time end of Jan).

Cheers!