[Cuis-dev] Syntax Error notification undo bug

[Eric Brandwein]

I didn't think about doing a subclass of the Parser, that's a good idea!
I'll try it and ask you if I have any issues. Thanks!

El jue., 6 jun. 2019 16:09, Hernan Wilkinson <hernan.wilkinson at 10pines.com>
escribió:

> Thank you Eric!
> You change is on github now.
>
> About making the "good" fix, sadly there are no tests for the parser,
> scanner or compiler... (at least that I'm aware of)
> The refactoring tests use the parser and compiler a lot but they do not
> cover all the paths.
> Making a change in the parser is difficult but not impossible, but be
> aware that you should do it using a subclass of the parser because if you
> make a mistake you wont be able to fix it because the parser does not work
> anymore :-)
> If you want to give it a try let me know and I'll show you how to do it.
>
> Cheers!
> Hernan.
>
> On Wed, Jun 5, 2019 at 3:22 PM Eric Brandwein via Cuis-dev <
> cuis-dev at lists.cuis.st> wrote:
>
>> Hi all,
>>
>> There's currently a bug when having an unrecognized selector in a method
>> and a syntax error at the same time. When the method is saved, and the undo
>> command is generated, the selector replacement and the syntax error
>> notification insertion of text are merged into one command, which in turn
>> makes the command not represent the actual user input.
>>
>> For example, if we don't have the m2 method defined, and we try to save
>> this method:
>>
>> m1
>>     self m2.
>>     [
>>
>> , an 'Unrecognized selector' popup will appear, prompting you to select
>> the correct selector. After selecting it, a syntax error text will appear
>> after the '[', and the source text will end up looking something like this:
>>
>> m1
>>     self m1.
>>     [End of block expected ->
>>
>> After undoing the last command, because the selector replacement was
>> merged with the syntax error replacement, the method will look something
>> like this:
>>
>> m1
>>     self m2ed ->
>>
>> I made a ChangeSet that fixes it by checking that the stop position of
>> the last replacement corresponds with the start position of the new
>> replacement before merging the two. But I believe the problem comes from
>> the fact that syntax error replacements should not be merged at all with
>> previous replacement commands; after all, user text replacements and syntax
>> error text replacements seem to me pretty separate undoable commands.
>>
>> Changing this without having tests for the TextEditor seemed like a
>> pretty dangerous task, so I didn't. Plus, the way I thought about changing
>> it involved changing the Parser, which seemed even more dangerous. There
>> probably are tests for the Parser, but I couldn't find them in the
>> BaseSystemTests. Where are they?
>>
>> Cheers,
>> Eric
>> --
>> Cuis-dev mailing list
>> Cuis-dev at lists.cuis.st
>> https://lists.cuis.st/mailman/listinfo/cuis-dev
>>
>
>
> --
>
> *Hernán WilkinsonAgile Software Development, Teaching & Coaching*
> *Phone: +54-011*-4893-2057
> *Twitter: @HernanWilkinson*
> *site: http://www.10Pines.com <http://www.10pines.com/>*
> Address: Alem 896, Floor 6, Buenos Aires, Argentina
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20190606/12e7531a/attachment.html>