RFC: remove traceback from "newer branch format than your bzr"

[Jelmer Vernooij]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Russel Winder wrote:
> I am with Emma on this.  And I seem to recollect we have had this
> debate in the past.
>
> For high quality programmers, there is no issue.  They implicitly
> understand the difference between internal and external problems
> and can likely differentiate them and deal with them.  Heck, they
> are probably doing exactly the same in the stuff they are writing.
>
> Weak programmers are just weak, and all too often incompetent.
> Sadly. We cannot ignore them though as they form the vast majority
> of programmers.  Sadly.  Let us treat them as non-programmers.
>
> More importantly, for me, are the real non-programmers who we are
> going to get using Bazaar, even if Bazaar is focused on programmers
>  as audience.  Actually I think in the end they will be the
> majority users.
>
> Non-programmers will be scared witless by the current penchant for
>  Bazaar to throw programmer informative stack traces.  This is
> especially true where the situation is an error and not a bug.
>
> I think that release versions of Bazaar should *never* display a
> stack trace.  bzr.dev and the RCs are a different matter since they
>  are being used in situations where the information is actually
> useful, by people who (generally) can deal with the situation.  No
> problem there.
>
> Stack traces in release versions are, for me, a huge problem,
> because they will frighten secretaries, managers, authors, and
> cause their bosses to be shocked and this will stop the spread of
> Bazaar.
Strack traces are the equivalent of segmentation faults in Bazaar,
when they are displayed there is *always* a bug, never expected
behaviour. I think hiding the detailed debug information here is a red
herring, we should work on fixing those bugs instead.

The stack traces provide a lot of useful information. I fear without
them we'll add another round trip to the interactions in bug reports
trying to extract them.

Cheers,

Jelmer

>
> On Tue, 2009-08-11 at 08:47 +0300, Alexander Belchenko wrote:
>> Emma Jane пишет:
>>> This evening I tried to grab a copy of Ian's new documentation.
>>>  I'm using the Bazaar that comes with Ubuntu (1.13) which
>>> happens to use the old formats. I was a greeted with a fairly
>>> scary message. It's at the bottom of this message.
>>>
>>> While I appreciate the usefulness of tracebacks, I'm not sure
>>> it's entirely necessary or appropriate in this case. More
>>> specifically:
>>>
>>> 1. The first line of the error pretty much says everything I
>>> need to know: bzr: ERROR: exceptions.KeyError: 'Bazaar
>>> repository format 2a (needs bzr 1.16 or later)\n' With the
>>> exception of the \n, I think this message does a pretty good
>>> job of telling me what the problem is. The rest of the message
>>> (all traceback?) could be omitted.
>>>
>>> 2. At the end of the message, which is what my terminal window
>>> scrolls after dumping a lot of garbage, I'm told to report a
>>> bug. I don't think it's a "bug" that I'm using an old version.
>>> Perhaps a new type of error message is needed?
>> No, in your case (big scary traceback) it's actually bug. So if
>> you'll file a bug about traceback on unknown format somebody will
>>  definitely improve this part, if not already done.
>>
>>> 3. Generally I find tracebacks really scary. I don't do
>>> anything in Bazaar that *should* generate errors (I typically
>>> work with fewer than 30 files and fewer than 500 revisions per
>>> project; I rarely do merges and use BzrUpload for most
>>> projects). I would prefer if the tracebacks were turned off by
>>> default with a message on how to enable them, but there's no
>>> way that I can see to turn them off.
>>>
>>>
>>> regards, emma
>>>
>>> PS I'm now using the PPA at Robert's recommendation and it's
>>> working well. Thanks!
> [ . . . ]

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
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=kkPw
-----END PGP SIGNATURE-----