Solved www/webkit2-gtk3 broken or synth not liking it?

PacketMan@ Even though I think it was poor form that you didn't even attempt to rebuild synth after explicit instructions (even after saying you wanted to "learn" FreeBSD), I'm going to remove any excuse you could have with disabling the watchdog. I spent my time to create 1-line patch for you.
  1. cd /usr/ports/ports-mgmt/synth
  2. mkdir files
  3. cd files
  4. fetch http://muscles.dragonflybsd.org/misc/patch-src_portscan-buildcycle.adb
  5. rebuild and reinstall synth
 
It's probably closer to the truth that you don't want to know.

Wrong again. When I have time I will dive in a learn it. I just really don't know that today. There are some basic skills there on that side of the OS that i don't have yet. I'm sorry.

PacketMan@ Even though I think it was poor form.....

Again I am sorry. I used evidence based facts (various commands) on my machine to try and show the statement of claims you were making. The output of those commands did not show that, and no one agreed with you or disagreed with me about what I said about those outputs. I specially asked for others to comment on the vmstat output and silence was regretful I say. I have very little free time in my life between working, volunteering, raising a family and some personal recreational activities (sport fishing), and thus I don't get to spend 40 hours/week on FreeBSD. I would love to get deeper into the OS and I will, but like my older machine it will be slow and take time.

I asked a simple question, and through a series of back n forths we determined the answer:
  • www/webkit2-gtk3 is not broken. It builds fine using portmaster.
  • Synth has no issue building other big long ports on the same slow machine.
  • And thus synth does have an issue building this port on slow cpu machines. (Remove the watchdog timer code and the issue will go away. You said so yourself I do believe.)
At the end of the day its up to you with the direction you want to take synth. I think the code is sheer brilliance actually, but with all due respect you do seem to get rather defensive with quick remarks before getting all the facts. If we (or you rather) are going to advocate the abandonment of Portmaster and replace it with Synth then Synth (in my opinion) really should work with the same machines that Portmaster supported. Often new users to the OS often tinker and learn using older hardware and we need to keep that in consideration. Advocating that users have multiple machines referencing one repository just to maintain the OS will be a show stopper for many newcomers to the OS. My opinion again yes, and go ahead and beat me it up, it doesn't matter to me. We will agree to disagree it seems.

Done with this discussion.
 
Done with this discussion.

I just want to point at that not only did I spend a ton of time discussing this with you, I also went an extra step and generated a patch for you and gave you additional instructions. Basic manners dictates a "thank you" at the very least would have been in order.

Wrong again. When I have time I will dive in a learn it. I just really don't know that today. There are some basic skills there on that side of the OS that i don't have yet. I'm sorry.

Did you attempt it? No? Even with step-by-step instructions? This explanation doesn't fly with me. The fact is that you didn't even try a very simple thing.

but with all due respect you do seem to get rather defensive with quick remarks before getting all the facts

Your conclusions that I labeled "snarky" I didn't agree with. You made some leaps there; they weren't "facts".

Advocating that users have multiple machines referencing one repository just to maintain the OS will be a show stopper for many newcomers to the OS.

Except that when you factor in that FreeBSD provides official packages and building your own packages on old hardware is not necessary, that blows this out of the water. And after all of that, you were given gift-wrapped a patch to resolve the issue of the ONE machine in evidence.
 
Okay guys.. Let's dial a notch down.

It's something PacketMan will have to solve on his own whether he took your advice or not.
 
Hello.
Sorry for necromancy, but I need to know a valid link instead of this outdated one.
I got the same problem, really underpowered machine and lang/gcc12.

Code:
/construction/xports/lang/gcc12/work/.build/./prev-gcc/xg++ -B/construction/xports/lang/gcc12/work/.build/./prev-gcc/ -B/usr/local/x86_64-portbld-freebsd13.2/bin/ -nostdinc++ -B/construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/src/.libs -B/construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/libsupc++/.libs  -isystem /construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/include/x86_64-portbld-freebsd13.2  -isystem /construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/include  -isystem /construction/xports/lang/gcc12/work/gcc-12.2.0/libstdc++-v3/libsupc++ -L/construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/src/.libs -L/construction/xports/lang/gcc12/work/.build/prev-x86_64-portbld-freebsd13.2/libstdc++-v3/libsupc++/.libs -no-pie   -g -O2 -fno-checking -flto=jobserver -frandom-seed=1 -ffat-lto-objects -DIN_GCC -fPIC    -fno-exceptions -fno-rtti -fasynchronous-unwind-tables -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wmissing-format-attribute -Woverloaded-virtual -pedantic -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings   -DHAVE_CONFIG_H -static-libstdc++ -static-libgcc  -o libgccjit.so.0.0.1 -shared \
     attribs.o jit/dummy-frontend.o jit/libgccjit.o jit/jit-logging.o jit/jit-recording.o jit/jit-playback.o jit/jit-result.o jit/jit-tempdir.o jit/jit-builtins.o jit/jit-spec.o gcc.o libbackend.a libcommon-target.a libcommon.a \
     ../libcpp/libcpp.a ../libdecnumber/libdecnumber.a  libcommon.a ../libcpp/libcpp.a   ../libbacktrace/.libs/libbacktrace.a ../libiberty/pic/libiberty.a ../libdecnumber/libdecnumber.a   -L/usr/local/lib -lmpc -lmpfr -lgmp -rdynamic  -lz  \
     driver-i386.o  \
      -Wl,--version-script,/construction/xports/lang/gcc12/work/gcc-12.2.0/gcc/jit/libgccjit.map  -Wl,-soname,libgccjit.so.0
gmake[5]: *** [/tmp/ccaXzPP0.mk:365: /tmp/ccIhF7Hf.ltrans121.ltrans.o] Killed
lto-wrapper: fatal error: gmake returned 2 exit status
compilation terminated.
gmake[4]: *** [/construction/xports/lang/gcc12/work/gcc-12.2.0/gcc/jit/Make-lang.in:168: libgccjit.so.0.0.1] Killed
gmake[4]: *** Deleting file 'libgccjit.so.0.0.1'
gmake[4]: *** Waiting for unfinished jobs....
###  Watchdog killed runaway process!  (no activity for 25 minutes)  ###



--------------------------------------------------
--  Termination
--------------------------------------------------
Finished: Sunday, 26 FEB 2023 at 09:33:35 UTC
Duration: 06:46:38

Thank you.
 
Sorry for necromancy, but I need to know a valid link instead of this outdated one.

Not a great idea though.
I got the same problem, really underpowered machine and lang/gcc12.
Just don't do that. Build on a machine that's suitable for the task, the "underpowered" one can just install the packages. If you don't have any, use official packages.
 

Not a great idea though.

Just don't do that. Build on a machine that's suitable for the task, the "underpowered" one can just install the packages. If you don't have any, use official packages.
But I need to do it in such way, just once. One little dirty hack to make it work would be acceptable.
 
But I need to do it in such way, just once.
What do you do on upgrades of that port?

When people write sentences like this, it's often (not always!) a sign of some XY problem. So, just maybe, there's a better way to solve whatever problem you're really trying to solve here?

I need to know how to disable watchdog or increase timeout in synth.
I just gave you the updated link to the relevant line in synth's source. I still warn you this doesn't sound like a sane thing to do.
 
What do you do on upgrades of that port?

When people write sentences like this, it's often (not always!) a sign of some XY problem. So, just maybe, there's a better way to solve whatever problem you're really trying to solve here?


I just gave you the updated link to the relevant line in synth's source. I still warn you this doesn't sound like a sane thing to do.
It's my home router, which is really weak in computation. I need to make some jails with custom build of infrastructure services, like DHCP + DNS + LDAP. I like synth and trying to make it on it's base.
Is it possible to make this option configurable?
 
It's my home router, which is really weak in computation.
Ok, but don't you have more powerful hardware that could build the packages? And ....
I need to make some jails with custom build of infrastructure services, like DHCP + DNS + LDAP.
... why "custom"? All these services are covered by binary packages, so, what exactly doesn't work for you?
I like synth and trying to make it on it's base.
I would prefer poudriere because it's the standard tool and therefore gets a LOT more support and activity etc ... but I guess synth should also be able to build a pkg repository that can be used on a different machine.
 
Ok, but don't you have more powerful hardware that could build the packages? And ....
No, I have just my laptop %(
... why "custom"? All these services are covered by binary packages, so, what exactly doesn't work for you?
I want to play around powerdns+ldap and dhcpd+ldap (both with SSL) options. Interesting to know how much work needed to make RFC2136 DNS updates possible for PowerDNS with LDAP backend. As far as I know this backend is just read-only. I didn't hear of any opensource DNS server with LDAP backend, except PowerDNS. Wanna give it a try...
I would prefer poudriere because it's the standard tool and therefore gets a LOT more support and activity etc ... but I guess synth should also be able to build a pkg repository that can be used on a different machine.
I'll do it as soon as possible.
 
Back
Top