ASMCA – Page 20 – Beyond The Data

rpm –target: Unkown option

today when Im tring to refresh kernel 2.6.39 on my Fedora11 node, some small issue hit during make rpm progress.

[root@suse-pc linux-2.6.39]# make rpm
make clean
...
...
rpm  --target i386 -ta ../kernel-2.6.39.ZIP.tar.gz
--target: Unkown option

[root@suse-pc linux-2.6.39]# make rpm

make clean

...

rpm --target i386 -ta ../kernel-2.6.39.ZIP.tar.gz

--target: Unkown option

double check and file the problem,

in the file linux-2.6.39/scripts/package/Makefile, this scriptwill detect the current installation of rpmbuild. If failed it will choose rpm instead. That’s why I hit the problem.

 23
 24 # Do we have rpmbuild, otherwise fall back to the older rpm
 25 RPM := $(shell if [ -x "/usr/bin/rpmbuild" ]; 
                   then echo rpmbuild; 
 26                else echo rpm; fi)
 27

24 # Do we have rpmbuild, otherwise fall back to the older rpm

25 RPM := $(shell if [ -x "/usr/bin/rpmbuild" ];

then echo rpmbuild;

26 else echo rpm; fi)

fix:
just refresh the rpm-build package and continue,.

yum install rpm-build -y

1	yum install rpm-build -y

[Recommend] upgrading wqy-zenhei font to V1.0-RC

讨厌升级的家伙又升级了一次小东西。

将正黑体的版本从V0.8升级到了V1.0-RC, 觉得最明显的进步是粗细粒度控制不错，尽管还有几只字型不满意，
但至少在字体放大缩小时候不会出现虚化锯齿现象了。

http://wenq.org/index.cgi?ZenHei

embr-r60 upgraded

必须承认我对于自己对于升级一切app的懒惰性。。。

embr似乎还用着0.2*版本,但zi.mu已经失效了一段时间了，却也没有其他要更新的地方。

在windows下着实不方面，用着默认的Terminal+secureCRT, ftp/ssh来搞定，

但上传好zip 解压时候发现问题，

...
...
  inflating: embr/img/ajax.gif
 extracting: embr/img/arr2.gif
 extracting: embr/img/bg-btn-green.gif
 extracting: embr/img/bg-btn-red.gif
 extracting: embr/img/bg-btn.gif      
         bad CRC 14ad5046  (should be fe7ac514)
file #44:  bad zipfile offset (local header sig):  30391
  (attempting to re-compensate)
file #44:  bad zipfile offset (local header sig):  30391
file #45:  bad zipfile offset (local header sig):  36930
file #46:  bad zipfile offset (local header sig):  37437
...
...

...

inflating: embr/img/ajax.gif

extracting: embr/img/arr2.gif

extracting: embr/img/bg-btn-green.gif

extracting: embr/img/bg-btn-red.gif

extracting: embr/img/bg-btn.gif

bad CRC 14ad5046 (should be fe7ac514)

file #44: bad zipfile offset (local header sig): 30391

(attempting to re-compensate)

file #44: bad zipfile offset (local header sig): 30391

file #45: bad zipfile offset (local header sig): 36930

file #46: bad zipfile offset (local header sig): 37437

...

仔细确认了一下，应该是ftp 没有换成binary mode,手动切了一下重传，一切ok.

更新基本无缝，很赞。不影响现有app~

setjmp and longjmp

I’m reading <Expert C Programming>

this topic 6.8 is on setjmp/longjmp mechanism,
compared with throw/catch this setting obviously has an re-enter callback.

notes,
-zip

#include &lt;setjmp.h&gt;

jmp_buf buf;

banana(){
	printf("In banana()n");
	longjmp(buf,1);

	printf("You'll never see this line.n");

}

main(){

	if (setjmp(buf))
		printf("back into mainn");
	else{
		printf("first time running n");
		banana();
	}

}

#include <setjmp.h>

jmp_buf buf;

banana(){

printf("In banana()n");

longjmp(buf,1);

printf("You'll never see this line.n");

}

main(){

if (setjmp(buf))

printf("back into mainn");

else{

printf("first time running n");

banana();

}

SIGLOST not support on Linux

这两天正好在看TLPI，有一处问题，正好给Mark写了一封信，人家回答也比较nice，不过更多么。。还是靠google了

SIGLOST This signal name exists on Linux, but is unused.

基本上我这边的Linux相关代码，都没有使用这个值，甚至一些架构(如MIPS)连comments这样的待遇都么有。

72 #define SIGIO 29
73 #define SIGPOLL SIGIO
74 /*
75      #define SIGLOST 29
76 */
77 #define SIGPWR 30

72 #define SIGIO 29

73 #define SIGPOLL SIGIO

74 /*

75 #define SIGLOST 29

76 */

77 #define SIGPWR 30

只有在Linux.SPARC上才找到这个信号的定义

76 #define SIGWINCH 28
77 #define SIGLOST 29
78 #define SIGPWR SIGLOST
79 #define SIGUSR1 30

76 #define SIGWINCH 28

77 #define SIGLOST 29

78 #define SIGPWR SIGLOST

79 #define SIGUSR1 30

当然，Solaris 10也有找到

 #define SIGLOST 37      /* resource lost (eg, record-lock lost) */</em></em>

1	#define SIGLOST 37 /* resource lost (eg, record-lock lost) */</em></em>

觉得比较有趣，记一下，顺便记一下Linux下 man signal(7)的描述

BUGS
SIGIO and SIGLOST have the same value. The latter is commented out in the kernel source, but the build process of some software still thinks that signal 29 is SIGLOST.