Re^5: XS efficient copy (SvCUR

The SvPV_nolen macro first checks the SV's private SVf_POK flag to see whether it contains a string. If it does, then it returns the pointer to the string, via SvPVX(SV*). If it doesn't, it calls sv_2pv_flags, which spits out a Use of uninitialized value warning, upgrades the sv, and returns (char*)"". That's safe to read from, but if you try to write to it... kaboom.

That was actually why I was calling SvPV_nolen (to force the SV to contain a string), but my understanding of it was a bit incorrect. I was expecting it to set that flag, but I was also expecting the return of the pointer to the existing buffer I had allocated with newSV, not an empty string.
When I realized that perl also had the length of the string (which, in retrospect, is obvious), the other part fell into place. I'm actually thinking that it IS safer to use SvPVX to get the buffer, so that I don't promote it 'by accident'.

Comment on Re^5: XS efficient copy (SvCUR_set)

Replies are listed 'Best First'.
Re^6: XS efficient copy (SvCUR_set) by creamygoodness (Curate) on Apr 09, 2006 at 15:35 UTC
I'm actually thinking that it IS safer to use SvPVX to get the buffer, so that I don't promote it 'by accident'. I agree that using `SvPVX` in this context is better practice. If you haven't allocated memory for the string, `SvPVX` will attempt to dereference a struct member that isn't there, and segfault. If you haven't both* allocated memory and made the string POK, `SvPV_nolen` will return a buffer that's not writeable, and you'll get a segfault when you try to write to it. perlapi seems to suggest that `SvPV_nolen` is safer. That is not true for this usage, and putting it in there lends a false sense of security. Better to use `SvPVX` -- so it's clear that you're working without a net. If you're lucky. In general, Valgrind is extremely helpful for detecting these kinds of problems, and I strongly recommend it. Sadly, I don't think it will help in this specific case, because it's Linux-only (there's also a FreeBSD variant, but it doesn't seem to work with Perl). -- Marvin Humphrey Rectangular Research ― http://www.rectangular.com	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re^6: XS efficient copy (SvCUR_set)
by creamygoodness (Curate) on Apr 09, 2006 at 15:35 UTC

I'm actually thinking that it IS safer to use SvPVX to get the buffer, so that I don't promote it 'by accident'.

I agree that using SvPVX in this context is better practice. If you haven't allocated memory for the string, SvPVX will attempt to dereference a struct member that isn't there, and segfault*. If you haven't both allocated memory and made the string POK, SvPV_nolen will return a buffer that's not writeable, and you'll get a segfault when you try to write to it*.

perlapi seems to suggest that SvPV_nolen is safer. That is not true for this usage, and putting it in there lends a false sense of security. Better to use SvPVX -- so it's clear that you're working without a net.

* If you're lucky. In general, Valgrind is extremely helpful for detecting these kinds of problems, and I strongly recommend it. Sadly, I don't think it will help in this specific case, because it's Linux-only (there's also a FreeBSD variant, but it doesn't seem to work with Perl).

--
Marvin Humphrey
Rectangular Research ― http://www.rectangular.com

[reply]
[d/l]
[select]