Tiny Core Linux
Tiny Core Extensions => TCE Bugs => Topic started by: andyj on February 27, 2016, 08:37:37 AM
-
@dentonlt: On my 32-bit platform submitqc requires --arch to check online repositories because my uname -m returns i686. Moving line 155 REPO_URL=... to after 180 (arch case) solves this problem. REPO_URL isn't referenced before line 180 so this shouldn't break 64 bit. Also adding
armv7l) BASELIBS="$BASELIBS ld-linux-armhf.so.3 linux-vdso.so.1"; ARCH="armv7" ;;
to the arch case should work for PI 2, but there isn't zsync or sstrip for armv7 for me to test it.
-
Hi, andyj -
Argh - browsing the forum feed and just came across your post. So sorry for the slow response - just hadn't seen it!
ARCH to line 180: totally agree. REPO_URL is wrong unless ARCH is defined, so this was a bug regardless.
Thanks for the ARM details - I've added these and will resubmit. sstrip is unneeded after all - strip is in binutils. I've also added a test to bail out on zsync creation if on ARM. I'm not a Pi/ARM user ... can you or someone else comment on whether the zsync file is necessary for the ARM repo, and how it's made if so?
Cheers ~
Sent from my HTC_0P6B6 using Tapatalk
-
Hi dentonlt
sstrip is unneeded after all - strip is in binutils.
You get smaller binaries using sstrip.
-
Good to know re sstrip vs strip. I'll use the former where possible, else fall back onto strip (arm7).
-
can you or someone else comment on whether the zsync file is necessary for the ARM repo, and how it's made if so?
zsync file is not needed to submit, at server side it is created automatically for all architectures.
-
there isn't zsync or sstrip for armv7 for me to test it.
zsync is part of base, see /usr/bin
-
Good to know re sstrip vs strip. I'll use the former where possible, else fall back onto strip (arm7).
sstrip.tcz added to 7.x armv6/armv7 repo
-
sstrip.tcz added to 7.x armv6/armv7 repo
Ah - done deal! Thanks, bmarkus.
-
can you or someone else comment on whether the zsync file is necessary for the ARM repo, and how it's made if so?
zsync file is not needed to submit, at server side it is created automatically for all architectures.
Thanks for this, too - I'll disable/cut that code in submitqc.
Sent from my HTC_0P6B6 using Tapatalk
-
I noticed that on system with i586 processor I get:
tc@box:/tmp$ /usr/local/bin/submitqc
submitqc: Unrecognized architecture 'i586'. See --help.
To make it work I have to use:
submitqc --arch=i686
-
Argh, this is a really old thread ... @neonix, if you're still out there and wondering ...
Short version: submitqc looks for your architecture (uname -m) and matches that to the available TC repos. Admittedly, i686 falls back to x86.
Thanks for mentioning - I think I can just have i586 fall back to x86, also ... but I have to follow-up on how that fits the --libs checks. I'll add it to the to-do list.
Cheers ~
-
HEllo Forum,
Think it goes
8086, ....286,....386,... 486DX... i586..... i686.... Pentium 1....., pent 2 ,...pent 3 ............ pent 4
ARm has 6 ,7 and 8's neon
thx
v
-
Hi dentonlt
... I think I can just have i586 fall back to x86, also ...
I see that:
. /etc/init.d/tc-functions
is only executed when -c/--color is passed to submitqc. That's unfortunate since tc-functions has functions to deal with
Tinycore specific issues. By including it you would change this:
[ -z "$ARCH" ] && ARCH=$(uname -m)
To this:
[ -z "$ARCH" ] && ARCH=$(getBuild)
Now submitqc would not require an update if other architectures are added.
There is also a getMirror() function that returns the full URL matching your system.