geda-developers team mailing list archive

[Peter TB Brett <peter@xxxxxxxxxxxxx>]

Hi folks,

I'm really sorry -- I may have caused gEDA 1.7.2 to ship in a completely
unusably broken state.  Two of my patches, that I thought were unrelated,
clashed in a way that breaks adding attributes!

I've now fixed it in git master.  See patch description:

>   gschem: Unbreak most of the dialog boxes.
>   
>   In commit 78c8ee905dad, OBJECT smobs were altered to store a TOPLEVEL
>   pointer for use when freeing them during garbage collection.  The
>   TOPLEVEL pointer is obtained using edascm_current_toplevel().
>   Unfortunately, since edascm_current_toplevel() raises a Scheme error
>   when the TOPLEVEL fluid is not set, this means that
>   edascm_from_object() can crash gEDA applications if it is not called
>   from inside a Scheme catch context *and* the TOPLEVEL fluid isn't set
>   in the current dynamic context.
>   
>   This exposed a bug in commit 71cc3eb5583b, where OBJECT smobs were
>   created *before* setting the TOPLEVEL fluid.  As a result, gschem
>   crashed pretty reliably when hooks were triggered via dialog boxes
>   rather than via keystrokes or menu item (e.g. when adding attributes
>   with the "Add Attribute" dialog box).  This patch fixes the issue.

I'm sorry for screwing this up.  :-(  There turned out to be a couple of
other bugs that break gschem a bit too...

I think it's feasible to put out a fast 1.7.3 release; the delta from 1.7.2
doesn't touch the libgeda .so at all.  Is that at all possible in the next
few days?

I also double-checked with several other users that `make distcheck`
*doesn't* choke on the Info file.  I'm fairly confident that there may be a
problem with Ales' setup that's causing his build failures -- I'm currently
waiting for him to get back to me about what the problem was that he saw,
so that we can debug it and work out what's going wrong.

Peter

-- 
Peter Brett <peter@xxxxxxxxxxxxx>
Remote Sensing Research Group
Surrey Space Centre