[Rd] Unfixed bugs in latest R-patched

Duncan Murdoch murdoch.duncan at gmail.com
Wed Jun 25 04:07:17 CEST 2014

On 24/06/2014, 10:26 PM, Radford Neal wrote:
>> Duncan Murdoch:
>> No, I don't think it's reasonable to expect you to write a patch, but
>> reporting the bugs in the R bug reporting system isn't that hard to do,
>> and does lead to fixes pretty rapidly in cases where the report contains
>> sample code to reproduce the problem.
> Sometimes.  Sometimes not.  For instance, PR #14985, a significant set
> of bugs, with an easy fix (patch provided), which took almost two
> years to make it to an R Core release - perhaps because you were more
> interested in trying to argue that they weren't really bugs at all...

I've just re-read that thread (and I'd invite anyone else to do so, it's
at https://bugs.r-project.org/bugzilla/show_bug.cgi?id=14985), and I
don't think that's a completely fair characterization of it.  That was a
case where problems did sit around for about a year before being dealt
with, but since it was a change to the interface that could affect third
party front-ends, I didn't put it into R-patched right away, which means
it waited most of another year before making it into a release.

>> "Fixed a problem in R_AllocStringBuffer that could result in a crash due
>> to an invalid memory access" sounds serious, but is just too vague to
>> follow up.  I would expect that doing a diff on the source files is
>> going to find all sorts of stuff:  pqR isn't just R with bugs fixed, it
>> has a lot of other changes too.
> You might expect that if it was really that difficult, I would have
> given more detail.  I think if you actually looked at this procedure,
> which is about 30 lines long, you might, seeing as you've been warned 
> that it has a bug, find the problem in about 30 seconds, even without
> looking at the pqR source code, which of course isn't difficult to do.

I think it's fixed now.

>> it would be more
>> helpful to the community if the bugs actually got fixed.  
> Indeed.
>> I think all of
>> the bugs that you reported last June got fixed within a couple of weeks
> Actually, the last six in the list I just posted are from the original
> pqR release last June.  Some of the six don't seem too crucial, but
> two of them seem reasonably serious (one leads to R crashing).

I imagine the reason they weren't dealt with was because we mistakenly
thought they had been.

Duncan Murdoch

>> Why not report them
>> more frequently than annually, and give the details you already have so
>> they are easier to fix?
> I did report at least one bug not long ago (which got fixed), after
> seeing (as now) that an R Core release was imminent, and therefore
> thinking it would be best if a fix was put in before it went out.
> You're of course welcome to look at the NEWS file at pqR-project.org
> whenever you like.
>    Radford Neal

More information about the R-devel mailing list