Re: [PATCH] coccinelle: api: add device_attr_show script

Re: [PATCH] coccinelle: api: add device_attr_show script

[Markus Elfring]

> +// Confidence: High

Would you like to add any suggestion for a possible patch message?


…
> +virtual report
> +virtual org
> +virtual context
> +virtual patch

+virtual report, org, context, patch

Is such a SmPL code variant more succinct?


…
> +ssize_t show(struct device *dev, struct device_attribute *attr, char *buf)
> +{
> +	<...
> +*	return snprintf@p(...);
> +	...>
> +}

I suggest to reconsider the selection of the SmPL nest construct.
https://github.com/coccinelle/coccinelle/blob/e06b9156dfa02a28cf3cbf0913a10513f3d163ab/docs/manual/cocci_syntax.tex#L783

Can the construct “<+... … ...+>” become relevant here?


Would you like to consider any further software design consequences
around the safe application of the asterisk functionality in rules
for the semantic patch language?

Regards,
Markus

Re: [PATCH] coccinelle: api: add device_attr_show script

[Julia Lawall]

[-- Attachment #1: Type: text/plain, Size: 1063 bytes --]



On Mon, 15 Jun 2020, Markus Elfring wrote:

> > +// Confidence: High
>
> Would you like to add any suggestion for a possible patch message?
>
>
> …
> > +virtual report
> > +virtual org
> > +virtual context
> > +virtual patch
>
> +virtual report, org, context, patch
>
> Is such a SmPL code variant more succinct?

This doens't matter.

>
>
> …
> > +ssize_t show(struct device *dev, struct device_attribute *attr, char *buf)
> > +{
> > +	<...
> > +*	return snprintf@p(...);
> > +	...>
> > +}
>
> I suggest to reconsider the selection of the SmPL nest construct.
> https://github.com/coccinelle/coccinelle/blob/e06b9156dfa02a28cf3cbf0913a10513f3d163ab/docs/manual/cocci_syntax.tex#L783
>
> Can the construct “<+... … ...+>” become relevant here?

<... ...> is fine if the only thing that will be used afterwards is what
is inside the <... ...>

julia

>
>
> Would you like to consider any further software design consequences
> around the safe application of the asterisk functionality in rules
> for the semantic patch language?
>
> Regards,
> Markus
>

Re: coccinelle: api: add device_attr_show script

[Markus Elfring]

>> +virtual report, org, context, patch
>>
>> Is such a SmPL code variant more succinct?
>
> This doens't matter.

Can less duplicate code be a bit nicer?


>>> +ssize_t show(struct device *dev, struct device_attribute *attr, char *buf)
>>> +{
>>> +	<...
>>> +*	return snprintf@p(...);
>>> +	...>
>>> +}
>>
>> I suggest to reconsider the selection of the SmPL nest construct.
>> https://github.com/coccinelle/coccinelle/blob/e06b9156dfa02a28cf3cbf0913a10513f3d163ab/docs/manual/cocci_syntax.tex#L783
>>
>> Can the construct “<+... … ...+>” become relevant here?
>
> <... ...> is fine if the only thing that will be used afterwards is what
> is inside the <... ...>

I propose once more to distinguish better if the shown return statement
may be really treated as optional for such a source code search approach
(or not).

Regards,
Markus