My findings about #519006
Hi,
I’ll be looking at debian-mips via GMane for a while, but I’m not
subscribed permanently so I’d appreciate being Cc’d.
As you can see, I have managed to isolate a testcase:
http://sourceware.org/bugzilla/show_bug.cgi?id=10144
→ http://sourceware.org/bugzilla/attachment.cgi?id=4550&action=view
I also managed, yesterday, on qemu (thanks aurel32 for the base
image) to build one of the packages with forcing -g off:
-rw-r--r-- 1 root root 1589 Jan 24 00:47 mopac7_1.15-1_mipsel.changes
To do so, however, I had to patch the cdbs *shudder* thing
in-place AND remove the -g from FFLAGS in debian/rules.
So I think, if squeeze comes close and the binutils bug
persists, we should consider changing the gcc-defaults
package to deliver a script instead of a symlink each,
which _ap_pends 「-g0」 to the user-provided compile
command, since patching the build system is often very
intrusive (as can be seen by cdbs *shudder* forcibly
adding -g – policy-backed, IIRC). Maybe make that de-
pendent on an environment variable, which the buildd
admin can set for packages affected by this problem.
On the other hand, let’s hope for upstream to provide
a fix or Doko(?) to find the changeset if it has been
fixed already, for now.
bye,
//mirabilos
--
Sometimes they [people] care too much: pretty printers [and syntax highligh-
ting, d.A.] mechanically produce pretty output that accentuates irrelevant
detail in the program, which is as sensible as putting all the prepositions
in English text in bold font. -- Rob Pike in "Notes on Programming in C"
Reply to: