kunta kinte

vcera jsem toho, stran placene prace, moc neudelal. v mailboxu na me koukal dalsi mejl z geni.com. pondeli, klasickej kopr.
tak jsem si rekl, ze se kouknu, co ze to tam sestrenice ze sedmeho kolena pridala.
a koukam, pridala tam nejblizsi rodinu okolo meho zatim nejzazsiho dohledaneho predka jana.
s dcerou sestrenice ze sedmeho kolena jsme se domluvili, ze odkazy na matriky budem davat do komentare. tak na nej kliknu.

9:30 - ty jo, ona fakt nasla dalsi generaci, stopy vedou do plasu

11:00 - plasy? koukam na historii plasu, tam snad nikdo po r. 1620 nebyl, krom jednoho ovcaka (kdovi jestli to nebyl zrovna muj predek). cas na obed.

12:30 - hmm, plasy jsou v nemcine plass, tohle vypada na platz, meli tehdy v chluncanech namesti? ale to preci neni namesti, to je ... FALC ... nasledujici hodinu mi euforie zablokovala mozek.

14:00 - euforie opada, ba primo mizi, kdyz mi dojde, co ze mi to nesedi na tom jmene. misto obvykleho kastner/kostner/khostner vidim kaschner. podezreni na falesnou stopu sili. jdu delat neco uzitecnyho. pisu sestrenici.

20:30 - decka spi, s chladnou hlavou prochazim 'gulas' v zaznamech. odstranuju duplikaty. kdyz si prohlizim dva zaznamy z r 1689 o umrti barbory a filipa zminujiciho bratra jana coby mistra polniho, dostavam napad - projit vse od data snatku (1635) az po prvni deti jana (1673). napad se vyplaci, nachazim vsechny inkriminovane osoby, plus minus par let vek sedi. nejstarsiho jana, pak barboru, mateje a nakonec filipa.

23:30 - konecne se odtrhavam od matrik, uz to neni euforie, je to radost. nasli jsme te kunto kinte :)


pred skoro 5ti lety jsem udelal mini-pruzkum v komunite 'rodokmen', kde jsem uvedl zminenou sekvenci jako spoustec meho hledani. otazek neubiva, spis pribyva. je myslivec probosta chotesovskeho klastera (jan kastner r. 1624 zminovany jako zesnuly) s adamem spriznen? sehral nejakou ulohu v jejich prichodu? odkud to vlastne prisli? falc je preci jen nejasny pojem...

dnf-plugins-core for epel7

with new mock are failing on epel7 builds for fedora 22 and above, because dnf builddep is provided by dnf-plugins-core, which is not in epel yet.

for those, who use mock on epel7 and needs build for fedora 22 and rawhide, there is https://copr.fedoraproject.org/coprs/jkastner/dnf-plugins-core/

build is based on 0.1.5 version, which corresponds to dnf-0.6.4, which already is in epel7.

udev rules - bInterfaceNumber

as owner of flyswatter2 (further fs2) i need very often use serial console of fs2, which is second ttyUSB interface.

problem is, that fs2 doesn't have always the same order.

i had already udev rules for ftdi devices to assign permissions to those ttyUSBn, but it was not enough.

as you can use SYMLINK="something", i wanted to have symlink /dev/ttyFlyswatter2 to be able simply attach to console using screen.

first what comes to mind is following:

ATTRS{idVendor}=="0403",ATTRS{idProduct}=="6010",ATTRS{serial}=="FS20000",ATTRS{bInterfaceNumber}=="01",MODE="0664",GROUP="dialout",SYMLINK+="ttyFlyswatter2"

but it doesn't worked. i found similar problem solution which i adopted to my needs:

ATTRS{../idVendor}=="0403",ATTRS{../idProduct}=="6010",ATTRS{../serial}=="FS20000",ATTRS{bInterfaceNumber}=="01",MODE="0664",GROUP="dialout",SYMLINK+="ttyFlyswatter2"

erasing mmc card in u-boot

i played little bit with beaglebone black and fedora rawhide images.
i was not able to boot from sd card, because of u-boot on emmc, even i used sd card boot button.
so i looked for something like 'dd if=/dev/zero of=/dev/mmcblk0', but in uboot to erase partition table on emmc and also MLO and uboot placed somewhere on first mb of emmc.
author of u-boot recommended me to use mw + mmc write. and some good hints regarding size of words for 'mw' and size of blocks which uses 'mmc write'.
this worked perfectly:

mw $scriptaddr 0 40000;mmc write $scriptaddr 0 800

but...

fedora's u-boot has by default as first mmc device sdcard, not emmc.
i somehow forgot this small detail.
and i was supprised, that i still see 'U-Boot 2014.04" instead of 'U-Boot 2013.10' after few resets and repeating command above.

is always good to check, which mmc device is active :)

/dev/loopN on fedora

for some reason loop module is not compiled in and systemd doesn't create nodes for loop devices.
if you need to mount/dump some partition on some image, let say on dump of mmcblk0 from embedded device, there is a way to do it:

• create loop device (must be repeated before any additional file processed by kpartx):

'losetup -f'

• create "looped" partitions

'kpartx -av some_dump_file'

• now you can do whatever you want on disk image and partitions like:

'sgdisk -p /dev/loopN'or
'dd if=/dev/mapper/loopNpNN of=dump_of_partition'

• when finished remove maps:

'kpartx -dv some_dump_file'

scp - ssh = brackets

i tried figure out how to transfer file from from laptop to ifc6410 board hooked to laptop via utp cable. using tcpdump i identified ipv6 address of ifc6410 board and connected using ssh.

connecting to remote machine is easy:

'ssh username@ipv6::ugly:addr:esss%ethN'

for scp that's not enough. scp needs brackets:

'scp filename username@[ipv6::ugly:addr:esss%ethN]:'

multiple tty opened in one screen session

while i was playing with 'console' as serial line i/o, i found nice feature:
'ctrl-a :' and typing 'screen /dev/ttyUSB0 115200' and then 'ctrl-a :' and typing 'screen /dev/ttyACM0 115200' i got 2 serial devices attached.
using 'ctrl-a "' i listed all screens:
###########
Num Name                                                                                                                                                   Flags

  0 ttyUSB0
  1 ttyACM0                                                                                                                                                  
###########