ldv changed the topic of #strace to: https://strace.io | https://strace.io/logs/ | strace 4.24 is out | strace-devel@lists.strace.io for dev discussions
Burgundy has joined #strace
Burgundy has left #strace [#strace]
sscox has joined #strace
pombreda has joined #strace
emachado has joined #strace
sscox has quit [Ping timeout: 250 seconds]
mjw has joined #strace
sscox has joined #strace
<mjw> ldv, that gold error doesn't immediately make any sense
<mjw> also, is there a #binutils channel?
<ldv> mjw: apparently, there is a #binutils channel with 3 members online
<ldv> mjw: there are also warnings mentioned by H.J.Lu
<mjw> ldv, apparently I didn't run the gold testsuite, sorry
<mjw> I think that is a bug in gold
<mjw> man, so much stuff gets compressed sections wrong
<mjw> It is too subtle
<mjw> Let me see if I can set things up so that it runs these gold tests.
<mjw> I admit I always get confused about this whole binutils-gdb setup.
<ldv> yes, binutils-gdb is somewhat messy
<mjw> I also don't like that they don't discuss patches at all
<ldv> that's because they believe that they don't introduce new bugs ;)
<mjw> that other fix hj just committed isn't wrong, but it could be more correct by initializing that variable only at the right place.
<mjw> It would make the code just a little bit more clear
<mjw> The commit message is also slightly misleading
<mjw> I do like the binutils testsuite btw.
<mjw> At least the parts that run by default :)
<mjw> It was really helpful. Wish I also got the gold part...
<ldv> Well, we run gold testsuite along with binutils testsuite
<mjw> sure, I just wonder why I didn't
<mjw> binutils could use a "o heay you just checked out this massive gdb-and-bfd-binutils-plus-gold-and-other-stuff source, here is how you properly configure and build stuff so that all proper testsuites run" documentation...
<mjw> aha, look --configure --enable-gold, now I get both your gcctestdir/collect-ld: internal error in do_write, at /home/mark/src/binutils-gdb/gold/output.cc:2063 and that warning about incorrectly aligned strings
<ldv> i wonder why --enable-gold is not the default yet
<mjw> that incorrectly aligned string warning thingy seems a simple bug in gold getting the alignment of the compressed strings wrong
<mjw> But that other thing is more mysterious.
<mjw> well, this is all a bit disappointing, gold has the same issue as bfd had
<mjw> it totally ignores the decompressed alignment
<mjw> ldv, this might take a bit, I have to grok the gold c++ coding style first. I think I understand what needs to be adjusted, but like the bfd fix it is not completely trivial. We need to teach the code that section alignment depends on the (decompressed) data alignment.
emachado has quit [Quit: Leaving]
pombreda has quit [Ping timeout: 250 seconds]
pombreda has joined #strace
mjw has quit [Quit: Leaving]
sscox has quit [Ping timeout: 250 seconds]