Re: Private Hebrew Web Site Template
On 4/26/07, Oded Arbel [EMAIL PROTECTED] wrote: I'm not familiar with typo2. I wouldn't touch typo3 with a laser guided radioactive stick. Very poorly written php, impossible to install. No documentation. -- Oded = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED] -- Cheers, Maxim Veksler Free as in Freedom - Do u GNU ? = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: New Israeli Debian mirror: archive CD images
On Tue, Mar 20, 2007 at 11:35:35PM +0200, Dmitry Sherman wrote: New Israeli Debian mirror: archive CD images Thanks a lot! http://www.debian.co.il/debian - Archive http://www.debian.co.il/debian-cd CD-IMAGES Make sure you publish it in the Debian's mirror list, with all the crypto signing stuff working (if you haven't done that yet)... V = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Groupware hosting
On 26/04/07, Danny Lieberman [EMAIL PROTECTED] wrote: Gil You must be kidding. Why in the world would you want a local ISP Run (do not walk) to Google Applications at www.google.com/a You have your own email, calendar and web site for free up to 50 users. I am running two businesses with it and I think it's outstanding. The mail is the familiar gmail UI and you do not get spam - period. The mail and calendar are integrated with none of this ridiculous and difficult to understand knee-jerk reaction to be Exchange compatible. And have Google have the right to do whatever they want with your business e-mail. Maybe you don't care but I know someone who's startup might interest Google and therefore he doesn't even use his Gmail account for anythinge relating to his work (e.g. he has a contract with Microfost. Other than that - I agree that the Google application are fantastic (I like the option to setup a location for an event and have it mapped with Google maps :)... Cheers, --Amos
producing video + presentations
Hi all, Sorry for being somewhat off-topic. I am looking, for my workplace, for a company that can come to a lecture arranged by us, with a presentation pre-prepared by the lecturer and with his cooperation, video it, and produce content for a web site that will have both, with some kind of automation, such that a user will be able to view/hear the video and the presentation will change slides in coordination. The content should work from apache without MS-specific extensions and work equally well on IE and FF, both linux and windows, using only very common plugins (e.g. flash). Of course the above is a kind of maximal wishlist and we are interested also in subsets. Also hiring professional individuals that have very good and relevant experience is an option, but we do not have enough work for a normal full (or part) time job. I am aware of the recent thread here discussing tools to achieve part of the above, and while interesting, we'd rather not do this ourselves. Please email me in private if you think it's inappropriate for the list. Thanks, -- Didi = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Tracing the functions stack through gdb
Problem: I wish to hack and solve Bidirectionality related bugs in Lyx. However, I don't want to read all and understand all the code in Lyx, but only the code related to cursor movement, character insertion, etc. How can I find the relevant pieces of code quickly? Suggested Solution: Run Lyx with a debugger, have the debugger print constantly which functions from lyx sources (so that printf() wouldn't litter the output) are being called, now press left, and watch the execution flow, based on that locate the required code. Is there any way to do that with gdb? Any other smart solution? = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Tracing the functions stack through gdb
Hi Elazar,
Put some fprintfs into the code, start with main().
Save the output to a file.
This is a quick way to circumvent all the gdb hassle.
This requires some code understanding.
Example:
== ce.h
#include cstdio
static char *cefilename = ce.txt;
#define ceprintf(...) \
do { \
FILE *dfp = fopen(cefilename, a); \
fprintf(dfp, __VA_ARGS__); \
fclose(dfp); \
} \
while (0)
End of ce.h ===
LyX main() fuction:
#include ce.h
..
main() {
...
ceprintf(Starting main()\n);
...
ceprintf(Calling somefunc or somemethod.\n);
someclass.somemethod(foo, bar);
ceprintf(back from somemethod.\n);
...
}
=
Check ce.txt file in your cwd.
Moshe.
Elazar Leibovich wrote:
Problem:
I wish to hack and solve Bidirectionality related bugs in Lyx.
However, I don't want to read all and understand all the code in Lyx,
but only the code related to cursor movement, character insertion,
etc.
How can I find the relevant pieces of code quickly?
Suggested Solution:
Run Lyx with a debugger, have the debugger print constantly which
functions from lyx sources (so that printf() wouldn't litter the
output) are being called, now press left, and watch the execution
flow, based on that locate the required code.
Is there any way to do that with gdb? Any other smart solution?
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
--
Moshe Gorohovsky
A6 CC A7 E1 C2 BD 8C 1B 30 8E A4 C3 4C 09 88 47 Tk Open Systems Ltd.
---
- Tel: +972.2.679.5364, http://www.tkos.co.il -
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
Re: Tracing the functions stack through gdb
Hi Elazar, Using the cpp macro that I have mentioned in my previous post eliminates the need to recompile all the LyX source files with debug information. No more g++ -g ! Elazar Leibovich wrote: Problem: I wish to hack and solve Bidirectionality related bugs in Lyx. However, I don't want to read all and understand all the code in Lyx, but only the code related to cursor movement, character insertion, etc. How can I find the relevant pieces of code quickly? Suggested Solution: Run Lyx with a debugger, have the debugger print constantly which functions from lyx sources (so that printf() wouldn't litter the output) are being called, now press left, and watch the execution flow, based on that locate the required code. Is there any way to do that with gdb? Any other smart solution? = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED] -- Moshe Gorohovsky A6 CC A7 E1 C2 BD 8C 1B 30 8E A4 C3 4C 09 88 47 Tk Open Systems Ltd. --- - Tel: +972.2.679.5364, http://www.tkos.co.il - = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Groupware hosting
Amos Separate paranoia from privacy. While I am skeptical that Google can afford to make use of privacy information or allow it to leak - if I were working on a startup that threatens search engines - I might also setup a private email system somewhere. FWIW - it is standard operating procedure these days for a customer to insist you use their corporate email. I have a contract with a US corp. and they insist we use their corporate (hosted Exchange) email system It helps them monitor outgoing content for data leakage. Danny danny On 4/26/07, Amos Shapira [EMAIL PROTECTED] wrote: On 26/04/07, Danny Lieberman [EMAIL PROTECTED] wrote: Gil You must be kidding. Why in the world would you want a local ISP Run (do not walk) to Google Applications at www.google.com/a You have your own email, calendar and web site for free up to 50 users. I am running two businesses with it and I think it's outstanding. The mail is the familiar gmail UI and you do not get spam - period. The mail and calendar are integrated with none of this ridiculous and difficult to understand knee-jerk reaction to be Exchange compatible. And have Google have the right to do whatever they want with your business e-mail. Maybe you don't care but I know someone who's startup might interest Google and therefore he doesn't even use his Gmail account for anythinge relating to his work ( e.g. he has a contract with Microfost. Other than that - I agree that the Google application are fantastic (I like the option to setup a location for an event and have it mapped with Google maps :)... Cheers, --Amos -- Danny Lieberman Software Security Specialists www.software.co.il - Secure your software www.opensolutions.co.il - Reduce operational risk www.controlpolicy.com - Stop data theft www.software.co.il/pta - Download a free copy of the PTA-Practical threat analysis tool Tel Aviv + 972 3 610-9750 US + 1-301-841-7122 Cell + 972 54 447-1114
Re: [off topic] Some new articles I wrote about science
On Wed, 25 Apr 2007 16:09:44 +0300 Uri Even-Chen [EMAIL PROTECTED] wrote: On 4/25/07, Shachar Shemesh [EMAIL PROTECTED] wrote: Completeness and Consistency relate to the relationship between the provability of an expression (syntax) and it's core truthfulness (semantics, or meaning). Since I was not talking about those, these hardly seem relevant. A theory cannot be either, because a theory is something that needs proof. In other words, using any moderately reasonable tools of proof, a theory can be correct and provable, correct and unprovable or incorrect (we usually do not let go of consistency because that leads to absurds). You will notice, however, that the theory is neither complete NOR consistent. These are measures not meant for theorems, but for logics. I agree. Even for logics, the statement above is incorrect. Zero order logic is both consistent AND complete. I'm not sure what zero order logic is. How do you say this sentence is not true in zero order logic? Trivial example: Is there a number of the form 2n+1, where n is a whole and positive number, that divides by 4? You answer seems to be: Sure there is (or, at least, you cannot prove there isn't). There are an infinite number of such numbers. My answer: No, there isn't. There is, indeed, an infinite quantity of such numbers, but ALL of them divide by 4 with a remainder of either 1 or 3. Good example. You assume this is true for all numbers. Take any positive number, multiply it be 2, add one, devide by 4, and you get either 1 or 3. Although I agree with you that it's true for any number we can represent by a real computer, I don't think it's infinitely true. I don't think integer numbers exist to infinity. We You think wrong, it's easy do define them to infinity can define numbers so big, that 2n and 2n+1 is almost the same. In You've got limits completely messed up. (2n+1)/2n does go to one in the limit but 2n+1 - 2n will always be 1 no matter how large the number n is any representation, whether in bits or in turing machines, if we devide both numbers in 4 we will not necessarily get two different Now you are talking about rounding errors and the problems of representation of floating point numbers on the computer not the results of exact arithmetic. BTW, as long as you are working with exact integers the results will always be 1 or three no matter how many bits you use (for bit represented numbers you are only interested in the bottom two in the case and you can have a zillion more that won't matter) The only thing that these mails show is some lack of understanding or knowledge of logic and what a proof actually is. You have a tendency to use points completely irrelevant to the problem to show that it is unprovable/unsolvable. Telling that the solution to the question to whether 10 is prime is dependent on what 10 is, is no proof to the question of how hard it is to compute it or whether it is computable. results. I can't define such a specific number, since you will be able to contradict me. It's an unknown unknown. Look what I wrote about the largest known prime number. http://www.speedy.net/uri/blog/?p=25 I'm referring to the general case, whether any specific problem is not in O(1). No specific problem can be proved not to be in O(1) for all algorithms. Ok. Let's take a simple one: Problem: You have a list of numbers in an array. You want to either create a new array, or in the existing array, list the numbers in reverse order to the input one. Proof: No number in the array, with the possible exception of the middle element in the case of an odd sized array, is located at the right place. Any algorithm, by the very constraints posed by the definition of the problem, must be moved. Ergo: The minimal complexity for an algorithm that performs this operation cannot be lower than O(n). QED It's not a decision function. Decision functions return either 0 or 1. I'm referring to the question whether there is any decision function which can be proved not to be in O(the size of the input). For example, if the input is a number represented in n bits, and the output is a sequence of its prime factors represented in m bits, then we are talking about m decision functions, each of them cannot be proved not to be calculated in O(n) steps. The total result would be calculated in O(m*n) steps, which is polynomial. But if each bit is calculated by a different computer, in parallel, then they can all be calculated simultaneously in O(n). If technology allows simultaneous calculations, such as neural networks, then maybe even O(1) can be achieved. Your example is simple - it can be calculated in O(n), and I just said that there is no proof that it cannot be calculated in O(n^2) [each bit in O(n), and for any specific input in O(1)]. Indeed, your function is a good example of calculating each bit in O(1) (on
Re: Tracing the functions stack through gdb
* Elazar Leibovich [EMAIL PROTECTED] [070426 16:24]: Problem: I wish to hack and solve Bidirectionality related bugs in Lyx. However, I don't want to read all and understand all the code in Lyx, but only the code related to cursor movement, character insertion, etc. How can I find the relevant pieces of code quickly? The best suggestion would be to ask the developers of LyX, some of them are also available on IRC and can easily point you to the best places to look at. Suggested Solution: Run Lyx with a debugger, have the debugger print constantly which functions from lyx sources (so that printf() wouldn't litter the output) are being called, now press left, and watch the execution flow, based on that locate the required code. I'd go for grepping the source with left right and cursor, as far as I remember this code was done in one large event loop so you should be able to find it from there, try to search for the above keywords in an area of a switch block. Baruch = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Groupware hosting
Danny Lieberman wrote: FWIW - it is standard operating procedure these days for a customer to insist you use their corporate email. I have a contract with a US corp. and they insist we use their corporate (hosted Exchange) email system It helps them monitor outgoing content for data leakage. wha? What would prevent me from using their hosted exchange for everything other than the actual sending of privileged information? What sort of a security system is that that only works if the attacker is willing to play by the rules? Danny Shachar -- Shachar Shemesh Lingnu Open Source Consulting ltd. Have you backed up today's work? http://www.lingnu.com/backup.html = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [off topic] Some new articles I wrote about science
On Wed, 25 Apr 2007 16:27:23 +0300 Uri Even-Chen [EMAIL PROTECTED] wrote: By the way, it's very easy to prove that for any specific decision problem, *there is* an algorithm who returns a correct answer in O(1). Consider these two algorithms: always return 0 and always return 1. At least one of them returns a correct answer for *any input*. For each _specific_ input. This version won't work for a class of problems and you can't say that you can build a table of answers since you still have the problem of choice, which taken straight forward is exponential (since the number of options grows exponentially with n). But if a general problem is considered to be hard, it must contain a subset of really hard problems, which cannot be calculated in O(1) by *any* algorithm. But at least one of my algorithms returns a correct answer for each of those really hard problems. Then a subset of really hard problems can never be proved to exist. And if a set contains only easy problems, then how hard can it be? Like I said, there are two O(1) algorithms, at least one of which is correct, but the choice between them can be hard, and you are interested in the choice, i.e in the correct solution, not whether it can be given easily once you know it for the _specific_ problem. This can even be extended to the halting problem. For any specific algorithm it is possible to return the correct answer in O(1), if a correct answer exists. If it is proved not to be computable in O(1), then a definite answer doesn't exist (any definite answer will lead to a contradiction). Like I said, you try to prove your point by answering a completely question than the one you posed. For a given problem there is a algorithm that gives the correct answer in O(1), you just don't know which one to choose. For a given class of problems (like whether an input integer is divisible by 4) the solution may or may not be O(1) since the one algorithm doesn't work any more, you need to chose one or the other and the question of choice is the issue and depends of the specific input. Uri. = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED] = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Groupware hosting
On 27/04/07, Danny Lieberman [EMAIL PROTECTED] wrote: Amos Separate paranoia from privacy. While I am skeptical that Google can afford to make use of privacy information or allow it to leak - if I were working on a startup that threatens search engines - I might also setup a private email system somewhere. Google is not just a search engine. And it's not about competing with - it could be simply having an idea of a service you can partner with them (or their competition in any of the numerous fields they operate in) which they can simply steal from your e-mail. FWIW - it is standard operating procedure these days for a customer to insist you use their corporate email. I have a contract with a US corp. and they insist we use their corporate (hosted Exchange) email system It helps them monitor outgoing content for data leakage. I'd imagine that the difference is that the host their managed exchange server didn't make them sign on an agreement which said that the host has the right to do ANYTHING with the files they store on that host's servers (including publishing it). Anyway - I've just read their privacy policy again and it looks like they improved it a lot - it now says that they won't access the content of your GMail messages (other than to provide the service) without express permission from you. So that point is apparently mute. --Amos http://www.google.com/mail/help/intl/en-GB/privacy.html
