Eclipse Community Forums: TMF (Xtext) » How to make Xtext parser stop if a particular word is found?

Help

Home

Home » Modeling » TMF (Xtext) » How to make Xtext parser stop if a particular word is found?

Show: Today's Messages :: Show Polls :: Message Navigator

How to make Xtext parser stop if a particular word is found? [message #723867]

Fri, 09 September 2011 18:06

Alex Ruiz

Messages: 103
Registered: March 2011

Senior Member

Greetings,

I have, IMHO, an unusual use case. There is an old version of the protocol buffer language that is very different to the actual (the open source) one. The problem is that the old and current version use the same file extension (.proto).

Since the old version is different that the current one, protobuf-dt (the Xtext-based editor for protos) shows syntax errors on files using the old syntax. There is no way to distinguish this two language versions from each other. Except for one keyword that is present in the old but not in the current language version.

Is there way to tell the Xtext parser something like "if you find keyword 'X' stop parsing and ignore any previous syntax errors"?

Many thanks in advance,
-Alex

Report message to a moderator

(no subject) [message #723962 is a reply to message #723867]

Sat, 10 September 2011 02:31

Henrik Lindberg

Messages: 2509
Registered: July 2009

Senior Member

Why not add the old syntax to the grammar to make it valid?
You can then detect if it is old or new syntax in the validator and flag
errors accordingly.

But maybe I misunderstood what you are trying to achieve...

- henrik

On 9/9/11 8:06 PM, Alex Ruiz wrote:
> Greetings,
>
> I have, IMHO, an unusual use case. There is an old version of the
> protocol buffer language that is very different to the actual (the open
> source) one. The problem is that the old and current version use the
> same file extension (.proto).
>
> Since the old version is different that the current one, protobuf-dt
> (the Xtext-based editor for protos) shows syntax errors on files using
> the old syntax. There is no way to distinguish this two language
> versions from each other. Except for one keyword that is present in the
> old but not in the current language version.
>
> Is there way to tell the Xtext parser something like "if you find
> keyword 'X' stop parsing and ignore any previous syntax errors"?
>
> Many thanks in advance,
> -Alex
>
>

Report message to a moderator

Re: (no subject) [message #725098 is a reply to message #723962]

Tue, 13 September 2011 23:48

Alex Ruiz

Messages: 103
Registered: March 2011

Senior Member

Thanks Henrik,

Adding the old syntax will make things go horrible wrong if a user accidentally uses syntax 1 mixed with sintax 2.

What I would like to do is to make the parser stop and ignore a file if one or more "words" are found. I hope there is a way to do this Smile

Cheers,
-Alex

Report message to a moderator

Re: (no subject) [message #725109 is a reply to message #725098]

Wed, 14 September 2011 00:23

Henrik Lindberg

Messages: 2509
Registered: July 2009

Senior Member

On 9/14/11 1:48 AM, Alex Ruiz wrote:
> Thanks Henrik,
>
> Adding the old syntax will make things go horrible wrong if a user
> accidentally uses syntax 1 mixed with sintax 2.
>
Since you have to discover this somehow, you could just a well discover
that in validation. Flag the mix as an error?

> What I would like to do is to make the parser stop and ignore a file if
> one or more "words" are found. I hope there is a way to do this :)
>

What is supposed to happen if a user tries to open an "old file". What
do they get in the editor?

Normally the editor is registered by file extension. You would need to
do this by examining content instead - i.e. parse the file to see if it
should be parsed. In order to do this, you need a parser (unless you can
do it with a simpler scanner).

Regards
- henrik

Report message to a moderator

Re: (no subject) [message #725416 is a reply to message #725109]

Wed, 14 September 2011 17:57

Alex Ruiz

Messages: 103
Registered: March 2011

Senior Member

Thanks Henrik.

The use case is not very usual, and it can be seeing as an odd one. The old syntax is valid, but:

1. We don't want to support the old syntax in the editor (as a way to encourage people to upgrade to the new version)
2. If users open a file with the old syntax, the editor should treat is as plain text (no syntax highlighting, outline, etc.)
3. Currently, if users open a file with the old syntax, the parser generates syntax errors that are inherited by imported files that use the new syntax. Users are complaining big time!
4. Both the old syntax and the new one have the same file extension!!

I'm not familiar with the solution you propose. I'll look into it in more detail.

Cheers!
-Alex

Report message to a moderator

Re: (no subject) [message #725418 is a reply to message #725416]

Wed, 14 September 2011 17:57

Alex Ruiz

Messages: 103
Registered: March 2011

Senior Member

I meant I'm not familiar with custom scanners and parsers Smile

Report message to a moderator

Re: (no subject) [message #725539 is a reply to message #725416]

Thu, 15 September 2011 06:37

Alexander Nittka

Messages: 1193
Registered: July 2009

Senior Member

Hi,

it might be possible to register your editor for a custom content type rather than the file extension alone. See the extension point definitions for
content type and editors.

The solution should rather be "open with the correct editor" rather than "let the editor handle wrong content".

Alex

Report message to a moderator

Re: (no subject) [message #725540 is a reply to message #725416]

Thu, 15 September 2011 06:39

Jan Koehnlein

Messages: 760
Registered: July 2009
Location: Hamburg

Senior Member

How about allowing the old syntax and the new one in parallel but adapt
syntax highlighting?

Am 14.09.11 19:57, schrieb Alex Ruiz:
> Thanks Henrik.
>
> The use case is not very usual, and it can be seeing as an odd one. The
> old syntax is valid, but:
>
> 1. We don't want to support the old syntax in the editor (as a way to
> encourage people to upgrade to the new version)
> 2. If users open a file with the old syntax, the editor should treat is
> as plain text (no syntax highlighting, outline, etc.)
> 3. Currently, if users open a file with the old syntax, the parser
> generates syntax errors that are inherited by imported files that use
> the new syntax. Users are complaining big time!
> 4. Both the old syntax and the new one have the same file extension!!
>
> I'm not familiar with the solution you propose. I'll look into it in
> more detail.
>
> Cheers!
> -Alex

--
Need professional support for Eclipse Modeling?
Go visit: http://xtext.itemis.com

---
Get professional support from the Xtext committers at www.typefox.io

Report message to a moderator