Hi Stuart, What do you think about a warning like:
!! Reading schematic form different directroy than the Setup file ./gafrc or ./gschemrc !! This is a warning for newbies as well as for old useres. They are warned that strange things may happen ..... And implemented in a view seconds!!! Later, when this is implemented, you and all the other hacker can play around with the code and try to find a 100% solution. I think this will take a while, because of the complexity of the problem. I expect a long testing time with instabable versions (what I dont like). Peter Am Donnerstag, 27. Januar 2005 03:08 schrieb Stuart Brorson: > Hi Guys -- > > Several newbie users have said that the symbol files were > misconfigured when they tried using gschem. In particular, they said > that locally defined symbol files didn't always work correctly. These > are symbol files pointed to by the local gafrc living in your > working directory. > > This is nonsense because the symbol files always install correctly. > Instead, I believe what is happening is that newbies sit in directory > A, and try to open a schematic in directory B. In the currently > released version of gEDA/gschem, this fails because gschem only looks > at the gafrc in directory A. It doesn't read the gafrc in direcotry B > so it can't find the sym files local to the schematic in directory B. > > I have fixed this by making a number of changes to libgeda. > In particular, gEDA/gaf now tries to open the gafrc living in the same > directory as your schematic. Everytime you open a schematic, it > checks if it needs to open the gafrc, and if so it opens it. > > These changes meant that I needed to also change gschem, gnetlist, > gattrib, etc. Specifically, I now call all the g_rc_parse stuff with > w_current as the first arg. w_current now holds a GList of all RC > files which have been visited, and if a new file is opened, the > corresponding RC file is also opened if it isn't already open. > > I also refactored code in all the programs which rely upon these > functions in libgeda. > > Question for Ales: I have tested this stuff on my box and it seems to > work. I would like to have others test it too. I can just upload it > to CVS, but it represents a lot of changes. Although I have tested > it, there is always the chance that it breaks something. So, hould I > just close my eyes and upload it, or would you like to look at it > first? Or perhaps put it in an experimental branch of the code? > > Finally, are there any volunteers who want to take the code out for a > long test drive? > > Stuart
