Peter,
It's really heavily influenced by the parser being used, with a handcoded
LL(1) parser you can include all sorts of fancy error processing and recovery,
The parser being used influences the technical problems of
getting thing to work well.
I have seen bad error recover with LL and very
Derek M Jones writes:
>Syntax error recovery was much better back in the 1960s and 1970 because job
>turn around time was so long.
>
>http://shape-of-code.coding-guidelines.com/2010/04/19/brief-history-of-syntax-error-recovery/
>
>When PCs became available compiler error recovery got seriously wo
lor...@gmail.com,
It has been interesting to watch the Rust community put real effort into
improving the error messages generated by the compiler:
Behind the bluster this is all about changing the format of the message.
Format is important because consistent layout reduces confusion.
Also poi
highlighting, inline errors (squiggles) etc.
Steven
From: ppig-discuss@googlegroups.com [mailto:ppig-discuss@googlegroups.com] On
Behalf Of lor...@gmail.com
Sent: 06 October 2016 05:25
To: ppig-discuss@googlegroups.com
Subject: Re: [ppig-discuss] Re: Beginner friendly error messages
On Wed, Oct 5
Hi Brett,
It's awesome to find more work in this area. Your thesis sounds really
interesting, please do email when it's up. I've downloaded your article and
am very much looking forward to reading it. I'm really glad to hear the
project is ongoing.
Please keep in touch!
Linda
On 6 October 2016
On Wed, Oct 5, 2016 at 10:13 AM, Derek M Jones wrote:
>
> The only way you are going to get user oriented error messages is by
> writing a front end designed for that purpose (which means ignoring
> the common "how to write a compiler" advice because this is not
> designed to produce meaningful m
Brett,
My work on enhancing compiler error messages in Java has been published by
*Computer
As a commercial compiler writer in a past life I know that error
messages rarely get much priority. Any programmer who knows what they
are doing should be able to figure the first message out (close,
>
>
>
>
>
> *From:* ppig-discuss@googlegroups.com [mailto:
> ppig-discuss@googlegroups.com] *On Behalf Of *Michael Sloan
> *Sent:* 19 May 2016 01:34
> *To:* Raoul Duke
> *Cc:* PPIG Discuss
> *Subject:* Re: [ppig-discuss] Re: Beginner friendly error messages
>
&
] Re: Beginner friendly error messages
Yup, this is a very astute observation. The deeper you are in a compilation
pipeline, the further you are away from the user's input. If there isn't
enough info relating the compilation back to the user's input, it is
challenging to giv
Yup, this is a very astute observation. The deeper you are in a
compilation pipeline, the further you are away from the user's input. If
there isn't enough info relating the compilation back to the user's input,
it is challenging to give helpful error messages. This can also be a big
problem for
> I'm surprised this isn't a solved problem.
The true state of the UX of programming (from error messages to language
syntax to tooling to ui to version control, to anything and everything) is
a clear indictment of the entire enterprise. Well, at least when it comes
to industry. For the large part
), Article ID
> 602570, 26 pages http://dx.doi.org/10.1155/2010/602570
>
>
>
>
> ------
> *From:* ppig-discuss@googlegroups.com on
> behalf of Thomas Green
> *Sent:* 06 April 2016 09:20
> *To:* Linda McIver
> *Cc:* PPIG Discuss
>
Hi Brett,
My apologies for the long delayed reply (to you and many others!). It's
been a crazy few weeks. Thanks so much for the references, that's a real
treasure trove. Now if you could email me the time to tackle it, all my
problems will be solved. ;-)
My current plan is to put it up as a hac
570,
26 pages http://dx.doi.org/10.1155/2010/602570
From: ppig-discuss@googlegroups.com on behalf
of Thomas Green
Sent: 06 April 2016 09:20
To: Linda McIver
Cc: PPIG Discuss
Subject: Re: [ppig-discuss] Re: Beginner friendly error messages
Hi Linda, good
Hi Linda, good to hear your voice.
>From way back, see Ben du Boulay's paper "Fatal error in pass zero" on
what not to say to novices : and Mark Eisenstadt co-authored a paper where
they looked at every error report issued to a class of beginners and
analysed every single one. That was in a very
15 matches
Mail list logo