[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA: pcb-20060422 snapshot

Daniel Wisehart wrote: 
On Sunday 23 April 2006 18:13, Ales Hvezda wrote:

Hi Dan,

[snip]


easy, there have been no "releases".  I've just periodically been making
snapshot releases which are releases in the sense that I verify that
'make distcheck' works for me which usually catches glaring errors in
the build system and makes sure that the snapshot tarball will have all
the files needed to build.  Also I've been branching the sources and
setting the version number to the snapshot date.  Other than that I have
not been doing anything extra.

Some day we might distinguish between true releases and snapshots but so
far we haven't.


	I haven't done a "release" of gEDA/gaf since, oh, 1998.  I sorta
like the whole dated snapshot approach, since the version number presents
some sort of a tie to reality (since it's a date, you know how old the
software is).  I haven't found any major problems/objections with this
approach.


The problem is that if a first time user picks up a snapshot that is broken--such as a PCB snapshot that won't load layout files--he probably isn't going to be using the tool for too long.

With gEDA, at least there is Stuart's wonderful installation ISO. It goes a long way to giving a new user an excellent initial experience. If they know to look for it there, Stuart has provided PCB as well, but Google searches for 'pcb install' or 'pcb release' did not list Stuart's website in the top 100.

My suggestion is that the PCB sourceforge page indicate that the snapshots given are alpha or beta snapshots and for a stable release they should see Stuart's ISO.

It is not at all clear to me that Stuarts ISO has a pcb that is any more stable than the snapshots. As far as I know he just grabs current CVS sources (which can really be broken) and makes a distfile. I'm not even sure that the version number is set to anything which would let us figure out what they're running.

I could be wrong but I think that recent screw up (my fault) with the file->load not working was not typical.

Anyone care to contribute regression tests, not sure how one does this with a gui program, that would be wonderful.

Part of the reason for a couple of quick snapshots in a row is we wanted to get people using the HID version to help shake bugs out of it and we have fixed a bunch of issues right away related to that. I'm hoping that things will settle down a bit now. Having anoncvs out has not helped with testing.

We just built the latest snapshot of PCB and it is broken with our configure options. Patches coming tomorrow. 

I'll be waiting.

-Dan