Richard Heck wrote: > Pavel Sanda wrote: >> José Matos wrote: >> >>> The next step is to guarantee that the conversion is correct and that >>> there are no data loss in the process. >>> >> >> it would be good to synchronize on this point once you start to work on >> it. >> currently i'm aware of two other relatedr tasks running which coordination >> would >> be good to not duplicate our work - one was my idea of some tool assuring >> that 1.x -> 1.x+1 file format transition produce identical latex output, >> >> > This is certainly impossible. Consider e.g. the subfloat->subfigure > transition in 1.5-->1.6. Of course, that's an exception to what we > generally intend.
i am aware of these issues. the main idea behind was that conversions which won't be equivalent would be manually accepted; we would have some kind of tree with documents which atomically represents features of lyx and if eg subfigure.lyx fails in 1.5 -> 1.6 transition one look in the output and if it is ok, than add some new output for 1.6 which would be considered in 1.6 -> 1.7 etc. in fact the idea behind was not originally to test lyx2lyx converison but to have kind of tool which would tell me what of _my_ documents needs attetion wrt output when moving from 1.5 to 1.6... but to extend this tool to lyx2lyx testing is easy job. pavel
