On Mon, 06 Aug 2001 08:12:41 GMT, "nicolas" wrote: > a �crit dans le message news: >> On Fri, 03 Aug 2001 11:31:42 GMT, "nicolas" >> wrote: >> >Beside that, the timestamp method cannot always be trusted. >> >We had problems with gnatmake which doesn't always recompile necessary >> >files. >> >> This so-called problem is a blatant example of fear-mongering ... > >Hum, is it really necessary to reply to this, I don't think so. Thanks for not digging a deeper hole. I will grant that you may have a bit of a problem expressing yorself in English, but your earlier statement does tarnish GNAT's reputation in providing a solid integrated make facility. What you claim is simply not true. > >> >The problem occurs when several files in different directories have same >> >name, same timestamp, and are selected with ADA_INCLUDE_PATH >> >timestamp and name of the file is checked, they are the sames, gnatmake >> >doesn't figure out that this is a different file and doesn't recompile >it. >> > >> >> ... because it looks like the file attributes have been dicked >> around with. To get similar symptons, I suggest you open the case on >> your workstation and pour soy sauce on your circuit boards. >> > >A very simple and normal use of cvs without any "dick around" shows the >problem. >Just try ... > I would never be caught using CVS in a closed developement environment. Boy, I feel the urge to merge; let's see I will grab this set of files that my colleague might be working on and then add some code. We will merge our way out of any mess we create. A process that does not trust gnatmake to do a bulletproof build, yet trusts wing-and-a-prayer merging techniques isn't worth arguing over.