On Wed, 09 Nov 2011 11:29:00 -0500, Daniel Macks <dma...@netspace.org> wrote:
> > On Sat, 5 Nov 2011 14:23:29 +0100, Par Kurlberg <kurlb...@gmail.com> wrote: > I was very happy to see that w3m-0.5.3-102 has made it into 10.7. > > However, if I run, say "w3m www.nytimes.com", I get a segfault (see > > below for what the window looks like.) "Amusingly", if I run under > > gdb (i.e., "gdb w3m" followed by "run www.nytimes.com") the segfault > > goes away. Any pointers on how to get it to work? (Will a bug report > > here reach the maintainer?) > > w3m compiled for me on 10.7 where it was a dependency of > can't-remember-what, but I guess whatever that was actually didn't > use it (or else runs a piece of it that doesn't crash). I just tried > it and it worked on a local helloworld.html and several remote sites > (both fetch and display) but crashed trying to access nytimes. Maybe > there's something noncompliant in their pages or a problem handling > some specific type of content? Here's an excerpt the crash-log > (~/Library/Logs/DiagnosticReports/w3m_...crash): > > Exception Type: EXC_BAD_ACCESS (SIGSEGV) > Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000008 > > Thread 0 Crashed:: Dispatch queue: com.apple.main-thread > 0 w3m 0x0000000103ed3a5b pushValue + 75 > 1 w3m 0x0000000103e8b8e5 flushline + 2549 > 2 w3m 0x0000000103e9008e HTMLtagproc1 + 1246 > 3 w3m 0x0000000103e93d95 HTMLlineproc0 + 1877 > 4 w3m > 0x0000000103e85d4f loadHTMLstream + 2047 > 5 w3m 0x0000000103e89e36 loadHTMLBuffer + 182 > 6 w3m 0x0000000103e839dd loadSomething + 29 > 7 w3m > 0x0000000103e884c6 loadGeneralFile + 6662 > 8 w3m 0x0000000103e73f9d main + 4989 > 9 w3m 0x0000000103e72af4 start + 52 > > Given your experience with gdb, I guess it would have to be debugged > with a bunch of added printf and recompilings? There's a -debug > option in "w3m -h" that says not to use it, but I did and it didn't > tell me anything other than that it crashed. I tried some of the > -dump* options, and it appears that the content is compressed > (Content-encoding: gzip ahead of a bunch of binary when using > -dump_extra). I uncompressed it and saved it to a local foo.html. w3m > still crashes the same way trying to render *that*. I tried -dump > -no-graph (to distinguish a parsing from a rendering problem) and it > hung (whereas worked correctly on helloworld.html). Definitely > something weird in the content of that specific page and poor > error-handling or sanity-checking in w3m's parser. One last note, that it does not fail on 10.6/x86_64 with http://www.nytimes.com (my failures are all on 10.7/x86_64; all my tests are using the latest standard xcode for each platform). dan -- Daniel Macks dma...@netspace.org ------------------------------------------------------------------------------ RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 _______________________________________________ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
