Re: Issue 2214 in lilypond: "fatal error" reported when version is too old but correct output is still generated

[-Eluze]

Graham Percival-3 wrote:
> 
> On Thu, Jan 12, 2012 at 02:29:05AM -0800, -Eluze wrote:
>> 
>> lilypond-4 wrote:
>> > This is deliberate to avoid misunderstandings in environments like  
>> > frescobaldi and related "lilypond IDEs".  You'll note that we create  
>> > the .pdf even though it reports an error.
>>
>> unbelievable! and even fatal!
>> 
>> that's the wrong way around - do you believe anybody is going to check
>> the
>> log anymore?
> 
> huh?
> 
> If we *don't* display an error, nobody on windows or osx is going
> to check the log.  That's precisely why we show an error, so that
> GUI programs will (hopefully) direct the user's attention to that
> log.
> 
> What's the big deal?  If this bothers you, just make sure you
> write the correct version numbers.  If you only use 2.14.0 syntax,
> then write 2.14.0 instead of 2.15.26.
> 

the word error should be reserved for severe situations, there are other
words to direct the attention to such discrepancies. for the actual "error"
a pointer, a note or another kind of message would suffice.

I often compare the result of the newest version with an older one and to do
that I would always have to change the version number.

and I also have an own environment which after compilation checks the log
for errors and warnings in different languages. if such are found the log is
displayed instead of the pdf/png.

so it would be nice to keep the words "error" and "warning" associated with
serious matters!

cheers
Eluze



-- 
View this message in context: 
http://old.nabble.com/Issue-2214-in-lilypond%3A-%22fatal-error%22-reported-when-version-is-too-old-but-correct-output-is-still-generated-tp33125536p33129135.html
Sent from the Gnu - Lilypond - Bugs mailing list archive at Nabble.com.