RFR(M): [ping!] 8139116: Fixes for warning "format not a string literal"

David Lindholm david.lindholm at oracle.com
Mon Oct 26 15:10:52 UTC 2015


Goetz,

This looks good in general, but I have one comment:

systemDictionary.cpp:

2014     size_t buflen = strlen(linkage_error1) + 
strlen(class_loader_name) +
2015       strlen(linkage_error2) + strlen(type_name) + strlen("\"") + 1;

This code is a little strange. I would prefer if you removed 
strlen("\"") (and replaced it with 1) and wrote a comment instead about 
this calculation.


Other that that, this is reviewed, but I'm not a 'R'eviewer. I can 
sponsor the change though.


Thanks,
David


On 2015-10-23 08:41, Lindenmaier, Goetz wrote:
> Hi,
>
> Could some Reviewer please have a look at this change?  I also need a sponsor!
> I fixed the issue reported by Dmitry and rebased the change to todays hs-rt:
> http://cr.openjdk.java.net/~goetz/webrevs/8139116-nonliteral/webrev.01/
>
> Thanks and best regards,
>    Goetz.
>
>
> From: Lindenmaier, Goetz
> Sent: Tuesday, October 13, 2015 11:32 AM
> To: hotspot-runtime-dev at openjdk.java.net
> Subject: RFR(M): 8139116: Fixes for warning "format not a string literal"
>
> Hi,
>
> The gcc build has -Wformat enabled.  This warns if the format string in a printf-like function is not a string literal, because if so it can not check that the types of the arguments match the format specifiers.
>
> In a row of places this warning is suppressed by PRAGMA_FORMAT_NONLITERAL_IGNORED. This change fixes all but three occurrences of this pragma.
>
> This especially improves heapInspection.hpp where a static buffer is used to print the format string, which is not multithreading safe.
>
> I didn't understand the purpose of the original code in heapDumper.cpp. Was this
> just bogus code, or did I oversee something?
>
> Also, I don't know why there was the workaround for '*' in heapInspection.cpp and globals.cpp.
> Is there any platform that does not understand this?
>
> Please review this change. I please need a sponsor.
> http://cr.openjdk.java.net/~goetz/webrevs/8139116-nonliteral/webrev.00/
>
> Best regards,
>    Goetz.



More information about the hotspot-runtime-dev mailing list