xalanc: update to 1.11

[mathew@…]

Xalan 1.11 was released in Oct 2012. Here is a patch file for the Portfile to build v1.11.

[mf2k (Frank Schima)]

Please attach a unified diff of the portfile so we can see your changes.

In the future, please Cc the port maintainer(s).

[jmroot (Joshua Root)]

• Please don't patch the Id line.
• The revision should reset to 0 (the default) when the version increases.
• Using an md5 checksum by itself should be considered deprecated. Using it in combination with another checksum type is OK if upstream publishes the md5. Looks like only a gpg signature is provided for xalan-c though, so after verifying the signature, you can just generate e.g. rmd160 and sha256 checksums to use in the portfile.
• What's the problem with clang? We'd much prefer to fix problems compiling with clang instead of blacklisting it, since it is the default compiler on OS X and the alternatives are getting more and more outdated.
• Please don't add trailing whitespace.

[mathew@…]

Thanks for the corrections... I don't generally make these, but had to update internally, so I thought I'd share the resulting Portfile.

• Please don't patch the Id line.

Fixed.

• The revision should reset to 0 (the default) when the version increases.

Fixed

• Using an md5 checksum by itself should be considered deprecated....

Verified signature and generated additional checksums

• What's the problem with clang?

There are no obvious problems with clang; Xalan builds just fine. Unfortunately, the resulting binary segfaults on a null pointer... presumably this is a compiler bug?

• Please don't add trailing whitespace.

Fixed

[ryandesign (Ryan Carsten Schmidt)]

Replying to mathew@…:

• What's the problem with clang?

There are no obvious problems with clang; Xalan builds just fine. Unfortunately, the resulting binary segfaults on a null pointer... presumably this is a compiler bug?

In that case, could you file a bug report against clang, and add a comment above the "compiler.blacklist clang" line with a link to that bug report, so that we can keep an eye on it?

What version of clang are you using? (clang -v) Have you tried any other versions?

[gallafent]

I patched the portfile using the Portfile.patch attached to this bug. I then removed the blacklisting of clang and built the port. The resulting executable runs successfully for my small testcase, no segfault.

My version of clang is (I think) current:

Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn)
Target: x86_64-apple-darwin11.4.2
Thread model: posix

This is running on:

Darwin Humber.local 11.4.2 Darwin Kernel Version 11.4.2: Thu Aug 23 16:26:45 PDT 2012; root:xnu-1699.32.7~1/RELEASE_I386 i386

(32-bit kernel since it's old hardware: I'm building for 64-bit:

/opt/local/bin/Xalan: Mach-O 64-bit x86_64 executable

)

I haven't done any more thorough testing ( ​http://xalan.apache.org/old/xalan-j/test/run.html#how-to-run-c appears both complex and outdated), but the simple smoketest I ran worked OK for me.

I therefore suggest that this portfile patch be applied but without blacklisting clang.

If there's an example test which generates a segfault for somebody else, I'd be interested to try that here with the version I built!

Definitely a good plan to get Xalan-C to version 1.11 though, 1.10 is very outdated!

[cooljeanius (Eric Gallager)]

Cc Me!

[kurthindenburg (Kurt Hindenburg)]

Cc Me!

[kurthindenburg (Kurt Hindenburg)]

done r142268