[RQt-devel] recent smoke update

[Michael Lawrence]

Unfortunately, I can't seem to reproduce the error, only the segfault. Btw,
your test code is off by 1 (so the error is happening on the first
iteration). It's also weird how the segfault is so predictable (it crashes
somewhere between 400 and 500 on my machine too).

Are you saying that this is a regression? I can't install the old version,
because it's broken for Qt 4.6.

Michael

On Sun, Jan 17, 2010 at 12:00 AM, Deepayan Sarkar <deepayan.sarkar at gmail.com
> wrote:

> On Sat, Jan 16, 2010 at 8:27 AM, Michael Lawrence
> <lawrence.michael at gene.com> wrote:
> >
> >
> > On Sat, Jan 16, 2010 at 7:33 AM, Deepayan Sarkar <
> deepayan.sarkar at gmail.com>
> > wrote:
> >>
> >> Hi (mostly MIchael),
> >>
> >> r206 (2010-01-14, Update to latest smoke) caused a regression in my
> >> setup because these lines were removed (they had been added following
> >> Justin's suggestion because I need the -fPIC flag):
> >>
> >> ----
> >> if( CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" )
> >>  set_target_properties(smokeqt PROPERTIES COMPILE_FLAGS "-fPIC")
> >> endif( CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" )
> >> ----
> >>
> >> Not a big deal, I just added them back, but I was wondering if there's
> >> anything we might do to prevent repeat occurrences. Was the file
> >> copied verbatim from the new smoke release?
> >>
> >
> > Yes, those files are copied verbatim with the sync. We need to get these
> > changes into upstream. But the uptake seems pretty slow. There is this
> fix,
> > as well as the fixes for OS X. We probably need to create our own branch
> of
> > the kdebindings stuff or at least maintain a set of patches. The latter
> is
> > probably a better idea.
>
> Yes, maintaining patches is fairly common practice, although I'm not
> sure how it's actually done in practice.
>
> Anyway, have either of you noticed more instability than there used to
> be? It's not very reproducible, but something always seems to go wrong
> after a while. In one case, the result of calling Qt$QPen() was
> reported as being of class "QGraphicsItem" (and there were some of
> those around), which may give a hint about what might be happening.
>
> Here's a small attempt at reproducing the bug (it's an infinite loop,
> so be ready to interrupt if nothing goes wrong). Notice how sometimes
> there is a segfault, and other times an error.
>
> --------
> dsarkar at sphulinga:tmp$ cat qbug.R
> library(qtbase)
> i <- 1L
> while (TRUE) {
>    i <- i + 1L
>    if ((i %% 100) == 0) message(i)
>    pen <- try(Qt$QPen(Qt$QColor(10L, 20L, 10L)), silent = TRUE)
>    if (is(pen, "try-error"))
>    {
>        message(i, ":", as.character(pen))
>        break
>    }
> }
>
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 100
> 200
> 300
> 400
>
>  *** caught segfault ***
> address (nil), cause 'unknown'
> aborting ...
> Segmentation fault
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 100
> 200
> 300
> 400
>
>  *** caught segfault ***
> address (nil), cause 'unknown'
> aborting ...
> Segmentation fault
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 100
> 200
> 300
> 400
>
>  *** caught segfault ***
> address (nil), cause 'unknown'
> aborting ...
> Segmentation fault
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 2:Error in assign(enum, structure(enums[enum], class = "QtEnum"), env) :
>  attempt to use zero-length variable name
>
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 100
> 200
> 300
> 400
>
>  *** caught segfault ***
> address (nil), cause 'unknown'
> aborting ...
> Segmentation fault
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 2:Error in assign(enum, structure(enums[enum], class = "QtEnum"), env) :
>  attempt to use zero-length variable name
>
> dsarkar at sphulinga:tmp$ R-devel --vanilla --slave < qbug.R
> 100
> 200
> 300
> 400
>
>  *** caught segfault ***
> address (nil), cause 'unknown'
> aborting ...
> Segmentation fault
>
> --------
>
> -Deepayan
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.r-forge.r-project.org/pipermail/qtinterfaces-devel/attachments/20100117/adf77300/attachment.htm