when you want to write a real program having all the features of speed, clean, memory management, multi-threading, maintainable, regardless small, big, linux, unix, windows, or when you want to open your creativity … or find a ready package to call, you directly start thinking of C.

but recently I started getting interest of JAVA, which is somehow a velvet C like language, free me to do things in a quick way, without any worries about memory management, and you can mildly mix toys together, to add wings to a fish

C++ I never liked although I uses for a long time, but it looks like C and JAVA Hybrid.

currently, I am not programming at all since long time, but still like to create my own tools, and I started to believe that any programming language goes away from 4th generation languages standards, and has no foundation libraries or classes, it will never succeed ( like BASIC), C succeeds always because you have everything you want, and you feel aways thousands of people thought of something before you and prepared it as a ready library (and very famous library even), just to learn and use.

when I write a program in C, I’ll not worry about the .Net or Java Runtime Environment is there on my friends/colleagues PCs or not, just send the binary/executable and do not worry about the rest, your program will be your man!

I am not trying to tell that C is perfect, I wrote some program that reads complex and different structures, in C it was very complex, and even I could not continue maintenance, so, I decided to transform my experience in writing this program to be done in OO, I shrunk the size of the program to 1/10 or the original one, and it is still maintainable.

I wish to find C language, and Java are mixed in C powerful and native compiler, and java cute way of dealing with OO, classes, and ignoring the stupid prototyping in C, and can call C previous libraries, and not to lose the powerful memory allocation or to access the byte as byte.

I did not meet a language that mixes between java way of using implicit pointers and declaring pointers when necessary.

finally, Google created a strange language called Go, I tried to use, but it looked to me something like a nightmare, although when I recognized the language features it was amazing, but I could not digest the way I have to write its programs, and when I came back to my program I could not remember what I was trying to do there!!

Finally C is the base of everything, it will keep so, until they invent different computing theory in the next 10,000 years.

I’ve been playing with your libumem module for dbx – but no joy, so far.

Here’s what I’ve got – a simple program that corrupts memory.

When I use mdb, I get this:

05:14 PM sundev19:c++ $ mdb a.out
> ::run
leaking 20 bytes…
leaking 20 bytes…
leaking 20 bytes…
leaking 20 bytes…
leaking 20 bytes…
mdb: stop on SIGABRT
mdb: target stopped at:
libc.so.1`_lwp_kill+8: bcc,a,pt %icc, +0×10
mdb: You’ve got symbols!
Loading modules: [ ld.so.1 libumem.so.1 libc.so.1 ]
> ::umem_log
CPU ADDR BUFADDR TIMESTAMP THREAD
1 0006a190 00155ef8 20ccc937bc8c0 00000001
1 0006a12c 00155f28 20ccc574506a0 00000001
1 0006a0c8 00155f58 20ccc1b139930 00000001
1 0006a064 00155f88 20ccbdf763e00 00000001
1 0006a000 00155fb8 20ccba3c47b10 00000001
0002a1f4 00000000 0 00000000

But when I run the same code with dbx and your module, the transaction log is supposedly empty!

Poking around inside dbx, I pulled up this data structure, which looks fishy to me – but I’m no expert.

What do you think?

(struct umem_log_header *) umem_transaction_log = 0×28008
(dbx 24) p * ((umem_log_header *) umem_transaction_log)
*((struct umem_log_header *) umem_transaction_log) = {
lh_lock = {
flags = {
flag1 = 4U
flag2 = ”
ceiling = ”
mbcp_type_un = {
bcptype = 0
mtype_rcount = {
count_type1 = ”
count_type2 = ”
}
}
magic = 19800U
}
lock = {
lock64 = {
pad = “”
}
lock32 = {
ownerpid = 0
lockword = 0
}
owner64 = 0
}
data = 0
}
lh_base = 0×2a000 “”
lh_free = 0×2000
lh_chunksize = 141733920769U
lh_nchunks = 1
lh_head = 0
lh_tail = 262144
lh_hits = 19800
lh_cpu = (
{
clh_lock = {
flags = {
flag1 = 0
flag2 = ”
ceiling = ”
mbcp_type_un = {
bcptype = 0
mtype_rcount = {
count_type1 = ”
count_type2 = ”
}
}
magic = 0
}
lock = {
lock64 = {
pad = “”
}
lock32 = {
ownerpid = 0
lockword = 0
}
owner64 = 0
}
data = 0
}
clh_current = (nil)
clh_avail = 0
clh_chunk = 0
clh_hits = 0
clh_pad = “”
}

)
}

Thanks in advance for your help!

The TOPLAS paper referenced (2008) give the NT-CSP. Email me (dcooke@coe.ttu.edu) and I can also send a good description.

D.C.

Send me an email and I will send you an intuitive definition of CSP-NT. The TOPLAS paper referenced is also a good source of this information.

-D.C.

I’m by no means an expert on either OpenOffice or Google Docs, but I’d say it’s an apples to oranges comparison..

OpenOffice was never meant to compete against Google Docs. OpenOffice was meant to be a free replacement for the MS Office suite.

I’m not surprised at all that Google has one-upped/raised the bar. Good for them, good for us, says I!