About the cpwiki . . .

I'm sorry if this is a bit out-of-date, but I read some of the articles and came across cpwiki.

Now, here's my problem. I'd like to contribute, but it says that I have to be one of the admin or editor groups.

Should this be fixed? Aren't Wikis supposed to be open and editable by all? (Or at the very least, require someone to sign up with Sourceforge first . . .).

Since when I click on the "login" link, it requires me to login with my Sourceforge username/password combo. Fine. I login, and I can't edit anything. (Presumably because I'm not part of the Sourceforge project. Fine. It tells me to "contact one of the project admins" to be added to the project.)

Is this not a direct violation of who can contribute to a wiki? Or does cpwiki not follow the Wikipedian rules?

Confuzzled me cannot comprehend this.

Thanks.

I just noticed this myself. I suspect that zacs7 upgraded on November 4th.

I'm looking into it.

I'm getting a password dialog box.

I was able to log in at the password dialog box by specifying my username and password from before, but I wasn't able to figure out how to create a new user. Nor was I able to edit any pages whatsoever once I had logged in, even my own user page. :rolleyes:

Sorry, I'm not sure what's up.

Yes I upgraded, because it was br0ken. However, I used the wiki that sf.net offered because the original was going to break with the sf.net upgrades. The accounts are done through sf.net.

> Is this not a direct violation of who can contribute to a wiki?
Not at all. Everyone is welcome to contribute you just need an account.

Everyone is welcome to contribute, however when it was open slather before the wiki was vandalized.

To login just change it to "https", you will need to be added to the project first however. Just pm me or the other admins on the project page (http://sourceforge.net/projects/cpwiki/) if you want to be added.

I'm not trying to control the wiki at all if it seems like that, I just fixed it so it wasn't broken and all previously referenced urls work. And this way we get automatic wiki upgrades :)

It's on sf.net because there was really no other place to host it at the time (nor does there seem to be now).

[edit]
Okay fishies, I *think* I know how it works.

Just your sf.net username and password to login, you'll then need an admin to give you the "editor" flag. You don't need to be a member of the cpwiki project (I think :)).

So list your username here (once you've logged in) and I'll give you the editor flag and we'll see if it works.

Nightowl (Nightowl256): I gave you the editor flag, see if it works.
[/edit]

Works for me now. I tried making myself an admin, and zacs7 did too, so something worked evidently. :)

I wonder if there is a way to make new users automatically added to the editor group, or allow the Users group to edit pages? I suspect that if we have to enable writing for users, we won't get as many users as we would otherwise.

I do agree. People are lazy to sign up, and even more annoyed if they must wait for permission.
Unfortunately, I cannot see a way to either allow non-editors to edit, or to change just registered people to the editor group.
The admin interface kind of sucks, in other words.
Well, for as long as an admin has to grant permission, I can grant permissions, as well, and I am on frequently, so a PM with the sf username and I'll move you to editor.

Still, this is a problem that really should be fixed... *grumble* *grumble*

Hey so are the editors/admins of the wiki still the same?

I have to transfer an existing web db interface -- 20000 pages -- to a (Semantic) MediaWiki, so for now I have mediawiki set up at home and before I design the automation scripts to transfer the material, I need to get a feel for how mediawiki works. Which means just writing some pages, I guess.

Of course, these can just be junk but I doubt I will get the same feel or insight that way, so I'm going to have to write real pages. So they might as well be real pages that I can cut and paste to somewhere else (such as cpwiki) later. I also don't want to bother doing wikipedia type stuff (find a new topic) or uncyclopedia stuff (just being coherent is fine, never mind funny), the best choice is probably some basic C stuff that people ask about like how to write timers or use pthreads or something.

And of course I think it's a good idea.

Anyway, no rush. I have a sf account but I can't edit or add a page.

Yeah, that's a known limitation with the current wiki. I'm still an admin, so gimme your account and I'll make it possible for you to add/edit pages.
Send me a PM in case you don't want to post it publically.

Alright gang, I added a few pages:

https://sourceforge.net/apps/mediawi...=Serialization
https://sourceforge.net/apps/mediawi...izing_an_array
https://sourceforge.net/apps/mediawi...e=POSIX_timers

That last one is my piece de resistance :p

For those of you unfamiliar with the philosophy behind wiki proliferation, if you write a page and leave a [[link]] in which does not exist yet, that link appears in red. Anyone (with editing privileges, PM Elysia) can then click on that link and start the new page. So if you are bored and short of ideas, here's a few of the topics I left red links to:

Structures
Linked Lists
Arrays
Preprocessor Defines
Error Checking
Networking
Function Pointers

Most of these could be adequately began with 5 minutes work. A page on "arrays" or "structures" may seem pointless, but it could easily become a repository for little known facts and details, sort of like how Epy was talking about his programming notebook a week or so ago.

If you want to add a new page not from an existing dead (red) link, you can create a dead link by listing the new page in the "annoted page" list on the main page and then use [[that]].

I can't recommending any online docs for mediawiki because I'm using a book, but I'm sure they exist and it is pretty simple. If you go for the "infer from existing code" method, since you are programmers, beware that this:

Code:

---

[[Arrays|an array]]

---

does not indicate that this will match one "OR" the other. The first part is the CASE SENSITIVE name of the actual existing page, the second is the term you wish to appear as a link to it. If they are one and the same you can just use [[Arrays]].

It might be nice one day to upgrade to MW 1.9 so we can install an extension to use code highlighting, but for now you will have to settle for <pre> tags.

Someone may want to write an article about clearing the input buffer. So common problem, yet I am too lazy to write one. Another thing to add to the list of someone interested.

Quote:

---

Originally Posted by Elysia

Someone may want to write an article about clearing the input buffer. So common problem, yet I am too lazy to write one. Another thing to add to the list of someone interested.

---

So you want to clear an input buffer? "Clearing an input buffer" is a [[stupid|clever]] thing programmers will say instead of "ignoring some text." Chances are you are reading this because you've witnessed what happens if you don't ignore sufficiently. Data gets misplaced. Let's discuss fixes.

The smart thing to do is avoid the problem by storing each line of text completely. In C and C++ there is a newline character, '\n', that marks the end of the line. It doesn't matter if you're getting input from the keyboard or a file, you can depend on its presence: The last thing a user will do before the program works with the input is press enter, so you know the input is complete, even if it's malformed or bad input; files will have '\n' at the end of their lines. It looks like this in code:

Code:

---

/* C code: */
char section[BIGBUFF];
char *line = NULL;

while(fgets(section , BIGBUFF , stdin) != NULL) {
  char *temp = NULL;
  if(line != NULL) {
      temp = realloc(line , strlen(line) + strlen(section) + 1);
  }
  else {
      temp = realloc(line , strlen(section) + 1);
  }
  if(temp == NULL) /* plan to exit with error */ break;
  line = temp;
  strcat(line , section);
}

/* If there were no errors we can use the string data as an argument to sscanf,
  or whatever you may need to do: */
if(ferror(stdin) == 0) {
  parse(line);
}

---

This is the store everything, worry about validation later approach, a winning approach. This approach is also the one you should use for text files 99% of the time. Ignoring parts of input files can lead to [[garbage in, garbage out]] and all the misery that comes with it.

But perhaps you are a student and for one reason or another are not allowed to take it. You may be putting data directly into variables via cin or scanf, or getchar. You can still use what you know about '\n' to your advantage.

Code:

---

/* C code: */
int yourAge = 0;
char junk = '\0'
char first = 'A';

printf("What's your age?\n");
scanf("%d" , &yourAge);
while((junk = getchar()) != '\n' && junk != EOF);

printf("What's your first initial?\n");
first = getchar();
while((junk = getchar()) != '\n' && junk != EOF);

/* Example output:
What's your age?
42 abc\n
What's your first initial?
\n
*/

---

Consider the answer to the first question. We know that 42 is interesting. Scanf will read that into the variable good enough, but the other stuff ought to be ignored. If we follow the same strategy of reading until '\n' (in effect, the while loop above), we can pretty much guarantee a fresh read from the keyboard every time.

The answer to the second question introduces us to a problem with our handy loop. If the first = getchar(); statement accepts an '\n', our handy loop will wait for the person to type more! You've started a fresh read by accident. Unfortunately, there is not a lot you can do about this. If you stick to one function to read everything, you can avoid this, because you should either need to clean up after the same function every time, or never.

Consistency is what makes this work so well. For further reading, know how your file API functions work, and how they break, to make better input routines:
*[[Scanf woes]]
*[[fgets]]
*[[feof]]

Practice!

...So instead of writing down a reply here, why didn't you just put it on the wiki!?!

Actually there are two pages on the topic:
SourceForge.net: Leaves data in input buffer - cpwiki
SourceForge.net: Fflush - cpwiki

Yes, but they're incomplete in this regard.

Quote:

---

Originally Posted by Elysia

...So instead of writing down a reply here, why didn't you just put it on the wiki!?!

---

Don't have an account.

All you could do was create one and send a pm so I could add you to the editor group... it's a free-for-all wiki, ya know. You can add whatever comes to mind.

Quote:

---

Originally Posted by Elysia

Yes, but they're incomplete in this regard.

---

This one is very good I think ;)

STDIN pitfalls

Quote:

---

Originally Posted by whiteflags

Don't have an account.

---

WAAAAHHHHHH! WAAAAHHHHH!

Pretty sure you can join sourceforge without having to have a project there.

If you do have a OSS project that you want free hosting for, it is not hard to get either.

Good one.
*bookmarks*