Re: Regression test analysis - 2.13.35

[Carl Sorensen]

On 10/3/10 5:39 PM, "Neil Puttock" <address@hidden> wrote:

> On 4 October 2010 00:19, Carl Sorensen <address@hidden> wrote:
> 
>> Is it a bug, or just different ordering of the output?  The warnings are
>> proper, aren't they?
>> 
>> Can't we just say the warnings have reversed order, so that the changes in
>> the regtest output can be ignored?
> 
> I think users will expect warning messages to appear in parsing order.
> 

I agree that it would be better to have warning messages show up in parsing
order, rather than translation order.  But if the error occurs in
translation, there's no way to make the warning messages show up in parsing
order, unless we force translation to occur in parsing order.  And I don't
think that's reasonable.  The two stages are deliberately decoupled.

> This is more important for markup, since the warning message doesn't
> tell you where in the file the error occurred.

Perhaps we need to find some way of giving better feedback on markup errors,
although I don't know what it is.

If we're going to call this a bug, I don't think it should be Critical
(release-blocking).

Thanks,

Carl