Herby, my ccomments were not meant to say you are not competent enough to fix Glorp. I know you have been active as the maintainer of Amber for quite a while now and know you are an experienced Smalltalker. So this is not an attempt to make you look incompetent or "unqualified". I just wanted to point out that * I think that if there is a bug in Glorp, it should be communicated to the maintainers in order to make sure the fix is making it into newer Glorp versions and from there to all dialects that have a port of Glorp (Smalltalk is too much of a niche to be able to stand more and more niche-ification of forks and stuff, esp. for such a central part as Glorp which are way too important to only be maintained by one or two developers - which they unfortunately are, at least to my knowledge) * I am not sure if anybody from Cincom is listening here looking for Glorp problems, so I saw/see the danger of "private" fixes / forks * I fixed a few bugs in Glorp in the past just to find out that the concept was correct but the place to fix it was wrong (or at least would not heal all related problems). Glorp is complex and it has lots of layers. It is a good example of the "avoid responsibility" concept that was once (what a coincidence) formulated by Alan Knight in an article named "All I've learned about object orientation I learnt from Dilbert" (or similar) - so I was gad Niall looked into these and gave me feedback as well as a "full" fix So I mainly ask you to post your fix and problem description to the Glorp Mailing list / Google group. It would be a pity if your fix is buried in some fork of Glorp. Joachim Am 16.08.17 um 20:02 schrieb Herby VojÄÃk:
Esteban A. Maringolo wrote:
Herby,
Please apologize I can't take the time to look into your code.
Np, if someone eventually will.
I granted you write access to <http://smalltalkhub.com/#!/~DBXTalk/Glorp/>, could you upload your latter version to it?
Uploaded.
Regards! Esteban A. Maringolo
P.S.: I did not gain understanding of Glorp innards in that small amount of time, of course, but I went by the intuition that at the place I made the fix, the normal (base class, Mapping >>) behaviour is to convert the right sides of the relation, and the only subclass which has it overridden, namely DirectMapping, seems to do the same thing (when looking at the code), just in an optimized way. Except, it does not convert, which I believe, strongly, to be a bug left out while doing that optimization.
Herby
-- ----------------------------------------------------------------------- Objektfabrik Joachim Tuchel mailto:jtuchel@objektfabrik.de Fliederweg 1 http://www.objektfabrik.de D-71640 Ludwigsburg http://joachimtuchel.wordpress.com Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1