kicad-developers team mailing list archive

[Jon Evans <jon@xxxxxxxxxxxxx>]

Just thinking about the impact between #3 and #4 (I don't think #1 or #2
are good options)

#3 - impacts everyone who has used both the old and new versions,
regardless of whether they use a VCS.  Impact is a minor annoyance (weird
dialog box, pops up once per schematic, that they probably don't care about)

#4 - impacts everyone from #3 who uses a VCS (fewer people than #3 I would
think).  Impact is a minor annoyance (file changed even if they open/close
it without trying to change it, once per schematic).  Some VCS users might
care, others might not.  For users that do care, we could publish a notice
(on the website, in the docs, etc) explaining what is going on.  Once they
look this up, they will know what is going on, and will likely not be
annoyed for each additional schematic, because they will understand that
their schematic files had invalid data that is now fixed (i.e. it's not a
*meaningless* change to the data requiring a VCS commit).

In both cases, the VCS users have to do a commit.  But in #3, they also
have to click "OK" on an extra dialog box for every schematic.

So, based on this, I would vote #4.

-Jon

On Thu, Mar 30, 2017 at 12:33 PM, Wayne Stambaugh <stambaughw@xxxxxxxxx>
wrote:

> I have found and fixed the bug in this bug report:
>
> https://bugs.launchpad.net/kicad/+bug/1677282
>
> This bug creates invalid schematic files which sets the component unit
> flag to 0 and breaks the netlist generator.  The fix itself is simple.
> What is not simple is what to do about the invalid schematic files that
> have already been created.  I added code to schematic parser to fix this
> but this creates a dilemma.  Technically the schematic is modified which
> begs the question, what to do next.  None of the following choices are
> particularly appealing:
>
> 1) Do nothing and leave the file in an invalid status until the next
> time the user saves the schematic.  This is probably the most convenient
> but what about broken schematics being used by versions of kicad prior
> to this fix?
>
> 2) Set the schematic modified status which will trigger a save warning
> when eeschema is closed even if the user hasn't made any changes.  We
> already do this with SCH_SCREEN::SchematicCleanUp() that gets called
> lots of places outside of schematic editing.  The problem with this is
> that the user has no idea why they are getting a save warning when they
> didn't actually change anything.
>
> 3) Set the schematic modified status and inform the user that there was
> an error in their schematic file that was repaired on load and requires
> saving.  This is the most informative for the user but reeks of nagware.
>
> 4) Silently save the corrected schematic with no user interaction.  This
> will undoubtedly make VCS users unhappy.
>
> As much as I hate nagware, I like unexpected save warnings when I
> haven't changed anything even less so I'm leaning towards option 3.  Any
> feedback would be appreciated.
>
> Thanks,
>
> Wayne
>
> _______________________________________________
> Mailing list: https://launchpad.net/~kicad-developers
> Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~kicad-developers
> More help   : https://help.launchpad.net/ListHelp
>