[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH] include error codes in error messages

From: Neels Janosch Hofmeyr <neels_at_elego.de>
Date: Thu, 07 Feb 2008 13:48:49 +0100

Stefan Sperling wrote:
> On Wed, Feb 06, 2008 at 05:13:28PM -0800, Jack Repenning wrote:
>
>> On Feb 6, 2008, at 1:21 PM, Branko ??ibej wrote:
>>
>>
>>> Another thing that strikes me is that the whole errror number, e.g. 200011
>>> in the example above, is quite and completely useless; because you never
>>> *see* that number in the code! You see some constant + 11, so this display
>>> is quite useless for searching for the cause.
>>>
>> I think we're looking for "pretty displays that provide the sort of info
>> naive users would push into google," more than "information-dense displays
>> that the primary committers can do amazing tricks with," aren't we? And
>> basically, the naive user is going to paste in the error line, whatever it
>> says (or some substring). So why do we need any representation at all of
>> the constant offset?
>>
>
> And come to think of it, why do we need a numeric representation at all if
> pasting the whole error string into a search engine will yield meaningful
> results? Will another numeric identifier really make that much of a
> difference for a user looking for the solution to a problem?
>
The point was that some error string messages have been changing because
they were
meaningless to the svn user. So, after changing the error message to
something
meaningful, one cannot find a solution googling that new error message,
obviously.

So the whole point of error numbers has been to be independent of the
actual error
messages, to be able to only google the error number.

Another question would be whether that error number actually makes sense to
be searched for: e.g. the error SVN_something_PATH_NOT_FOUND could be raised
due to various reasons, and there wouldn't be one catch-all solution...

Someone ought to do some research whether using these numbers will at some
point in time improve the error googling success rate. Also considering
that right
now the error numbers aren't googlable yet, either, it might just make more
sense to simply change any error messages to something meaningful and wait
for that message to be googlable a month or two down the road.

I guess, in the end it probably won't make much of a difference either
way. So I'd
favor the non-error-number better-looking error report (considering that in
maintainer-mode the error numbers are displayed anyway).

neels

-- 
Neels Janosch Hofmeyr
Software Developer
neels_at_elego.de
Public Key at
 subkeys.pgp.net
or
 http://binarchy.net/neels/neels.hofmeyr.public.key.asc
elego Software Solutions GmbH           http://www.elegosoft.com
Gustav-Meyer-Allee 25, Gebäude 12       HRB 77719
13355 Berlin, Germany                   Amtsgericht Charlottenburg
Tel.: +49 30 23 45 86 96                Sitz der Gesellschaft: Berlin
Fax:  +49 30 23 45 86 95                Geschäftsführer: Olaf Wagner

Received on 2008-02-07 13:49:02 CET

This is an archived mail posted to the Subversion Dev mailing list.