Front page | perl.perl5.porters |
Postings from August 2016
On Wed Aug 24 12:13:58 2016, dcollinsn@gmail.com wrote: > Some diagnosis: > > Perl_yylex maintains up to two pointers, `s` and `d`, into the parser > buffer at PL_bufptr. It can call skipspace(), which can potentially > grow (and realloc) its argument. This can leave the second pointer > pointing at the old buffer. Under most cases it isn't visible, because > the old buffer isn't reused or zeroed. However, under Valgrind or > libdislocator, this memory management error becomes visible. > > Ideally, these would both just be offsets relative to PL_bufptr rather > than pointers, but I understand the desire have them be pointers for > performance reasons. This would involve refactoring Perl_yylex as well > as changing how skipspace is called (argument and retval would be an > offset against PL_bufptr instead of a pointer into PL_bufptr). > However, even just looking at skipspace, I don't understand this code > well enough to do anything like that. > > The simpler fix is to patch the holes by ensuring that the second > pointer is always updated when we call skipspace, as in the attached. > That fixes all of my testcases, not sure if Brian has any similar > ones. This also `make test`s clean. Both hunks look correct to me. (I have a guilty feeling I may have written one of those bits of code.) Could you add tests too, so that ‘make test.valgrind’ (or whatever it’s called) will catch any regressions? -- Father Chrysostomos --- via perlbug: queue: perl5 status: open https://rt.perl.org/Ticket/Display.html?id=129069Thread Next