Uwe Stöhr wrote: > Am 25.10.2015 um 11:05 schrieb Georg Baum: > >> I reverted to get rid of a regression you introduced. If you want more >> details, please read the commit log and the thread I referred to in the >> commit log, since we discussed the possible alternatives in detail over >> the last days. > > Well, of course I read the thread carefully but I still don't understand > where the regression is/was. What exactly did not work with my code but > worked before?
http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg189325.html > The alignment within a box on the list before it went in. I also > mentioned that the alignment is not yet shown within LyX and nobody > objected. This could have been added any time later. I found it now. The feature is mentioned in one single line in http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg187298.html, and the FORMAT entry in the patch does not mention it. Therefore it was quite easy to overlook it, but I admit that the problems could have been detected if somebody did read the patch really carefully. It seems that nobody did that unfortunatly. However, I am still convinced that the drawbacks would have been noticed if you had written about the problem you wanted to solve (the usual paragraph alignment produces unwanted space for certain boxes), and how you wanted to solve it (adding a horizontal alignment parameter to these boxes) before you started working. For the future please do that before you invest too much time in an implementation, this can really save time. > Scott stumbled over the topic because a bug in my lyx2lyx code. I find > it strange that I did not get the chance to fix this. Since when do we > revert when there is a lyx2lyx or in general when there is a regression > bug in a feature? We do not revert in general. We discussed the different options, and the result of the discussion was to revert _in this particular case_. >>> I sent you a test file to make clear what I wanted to achieve with the >>> box alignment and that I cannot get this via paragraph alignment. So how >>> can I now get this to work after your reversion? For now I am now forced >>> to use ERT but the idea was to get rid of it. >> >> Yes, you have to use ERT for now, which is a pity, but the best >> compromise we could find. Nobody volunteered to implement the proper >> solution up to now. > > What is proper in your opinion? I still don't get your point. I played > with your testcase and cannot see the problem. http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg189473.html A proper solution is one that - does not confuse the user by giving him the possibility to specify almost the same setting in two different places with very surprising results if both are mixed (look at whether the box or paragraph setting wins for each combination) - is implemented in a way which is consistent with how the same problem is solved for other insets > And as I wrote using the > paragraph alignment does not help for the alignment inside boxes as my > example showed. I know that (see above: "Yes, you have to use ERT for now"). Georg
