OpenEuphoria

1. The Archive and version compatability

• Posted by jeremy (admin) Mar 25, 2009
• 947 views

Since we are going to use the description for version compatability, can we all do our best to include something searchable? I am up for suggestions on what we should use, but I'll start with a few ideas:

---

REQUIRES: 4.x

This has a keyword to indicate that it REQUIRES something, however, when searching the archive right now for "requires 4" I get all sorts of things, anything with the word REQUIRE and the number 4. For instance, here's the first match:

WIN Binder GUI Interface 210K Rod Damon Nov 7/05  1.00    
A Windows GUI interface for binding and shrouding with  
Euphoria 2.4 & 2.5. Has been tested on WIN98 & XP for  
Windows & DOS. Written using Andrea Cini's EuWinGUI library. 
Requires the 2.4 or 2.5 RDS registered binder/shrouder  
(not included).   

This clearly does not require 4.x.

---

REQ40

This will be a good match, but pretty hard to read when put in with other text, but I am afraid we are going to have to go with something like this. Now, a further problem is going to be how to indicate something works on 2.x, 3.x and 4.x. We need to indicate that (as opposed to just leaving it blank) because The Archive has a ton of packages in it that will probably never be updated for 4.0 and most have no type of version indication.

---

Maybe we should have another keyword that we use such as COMPAT40, COMPAT31, COMPAT25 or something like that. So, a description may be:

This is my pacakge that does something good. 
 
COMPAT25, COMPAT30, COMPAT31, COMPAT40, COMPAT41 

---

If we fail to do something like this, then the archive will not be searchable. For instance, I want to search for any advanced math library that I know works with 40. My search term would be: math COMPAT40

---

Just thinking aloud on how to solve this problem. I'm very afraid that even if we agree on a common term that the archive is going to be pretty messy to find anything in because we will all forget, or mistype a character or simply not know, but we must do our best with what we have.

Thoughts?

Jeremy

3. Re: The Archive and version compatability

• Posted by DanM Mar 25, 2009
• 931 views

Maybe it should instead be "RequiresVersion", and then the searcher would still have to see which versions show up as a result?

Or, "Requires VersionX", then search for "VersionX"?

Dan

5. Re: The Archive and version compatability

• Posted by DanM Mar 25, 2009
• 942 views

Yeah, I think that makes better sense. And even with concerns with misspelling, maybe "Compatible VersionX" would be better? "Compat" just looks weird.

Dan

7. Re: The Archive and version compatability

• Posted by jeremy (admin) Mar 25, 2009
• 939 views

Bernie, There are plenty of things in the archive that work with 2.5, and some people may still have 2.5 installed on their system. Further, we do not have control of the archive, we cannot move anything. Even if it were moved, that would require changes to the archive script which RDS doesn't want to do right now. As for the special compatibility list or marking for 4.0, that too would require changes to the script. So, we are trying to figure out how to work within the confines we have been given.

Now, it's very possible to have code that works on 2.x, 3.x and 4.x w/o issue, so we need some way of indicating that as well.

Jeremy

9. Re: The Archive and version compatability

• Posted by jeremy (admin) Mar 25, 2009
• 920 views

I am not at all suggesting that we mark previous entries or even try to categorize previous entries in any systematic way. What I am suggesting is code that we are writing today, tomorrow and the next day have some type of standard tagging on it. I have written code that works in 4.0 and 3.1. If it were of general use and I upload it to the archive tomorrow, I want to be able to specify that it works in 3.1 and 4.0 with some type of common tag.

Now, what was discussed in the past was if you downloaded a library from the archive that looked nice and it didn't work on 4.0, then you should have been able to click a button on the entry in the archive that said "Does not work in 4.0" or something to that effect, so that others would know as well, but that idea has been canned as the the archive script will not be updated.

I am in no way suggesting anyone go back and try to update everything or even mark everything. I am specifically stating about from today forward.

Now, who is going to answer questions on converting code from 3.x to 4.0? I hope everyone on the forum. We all have code in 3.1 that is going to need to be converted. Not from the archive at all, our own personal code. I would venture to say we are all going to get pretty good about updating 3.x code to 4.0.

Jeremy

P.S. The \\ in your messages is not required and should not be used. Just type away, the forum will word wrap the text for each persons browser width. When you want to start a new paragraph, leave an empty line between the previous sentence.

11. Re: The Archive and version compatability

• Posted by jeremy (admin) Mar 25, 2009
• 956 views

Actually, that's the point. I downloaded DOT from Daivd Cuny, and it worked fine on 4.0. It was last updated in 1998! Now, it's a pre-processor and therefore doesn't support some of the new language constructs, but it ran fine and was last modified in 1998. I am betting that we will find all sorts of things like this that run in 4.0 w/o issue. With DOT I didn't have to change one single thing.

Jeremy

13. Re: The Archive and version compatability

• Posted by jeremy (admin) Mar 25, 2009
• 959 views

Rob would like to keep the archive on his site. They enjoy maintaining it.

Jeremy

15. Re: The Archive and version compatability

• Posted by Kat Mar 25, 2009
• 971 views

That's kinky.

But does that preclude having a detailed search facility on another site? Because i have openeuphoria.com and this Tiggr bot which already has the archives un-zipped and indexed, you see....

useless

17. Re: The Archive and version compatability

• Posted by Kat Mar 27, 2009
• 950 views
• Last edited Mar 28, 2009

Sorry to say, but due to an incident on irc, i won't be putting the db access online. I deleted the entire dir off the host, a host i just paid for.

I also won't be completing http.e, or submitting any other code. This will make more than one person happy as h, i know. I refuse to be drawn into more defensive positions. You want it?, you code it, you fight for it.

useless

19. Re: The Archive and version compatability

• Posted by Kat Mar 27, 2009
• 926 views
• Last edited Mar 28, 2009

A valid on-site header trumps no header. It's not about spoofing anything with malicious intent. It's not about malformed header lines. And sending the host address as the referer: tag does no harm. There is code to allow the http.e user to set the line to anything desired (except "", currently), just like setting the user-agent: line. I offered to code to allow "" be set, so the referer would not be sent, but that was declined.

I point out that using http.e with a user-agent line of anything other than http.e is worse than specifying the host name as referer.

I've run across domains before that allowed page views (yes, that is a "download") *only* if refered to by another page on their domain, which meant that bookmarking the desired page didn't work, shortcuts didn't work, offsite references by other hosts didn't work, the domain in question missed one or two web ad hits, etc etc etc.. Making life easier by sending the domain name of the host you are accessing hurts no more than http.e having :

public procedure set_sendheader_useragent_msie() 
	set_sendheader("User-Agent","Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0)") 
end procedure 

in it, or me coding

{"User-Agent",": ","Opera/5.02 (Windows 98; U) [en]"}, --</joke> pick your own  

Frying my tail over something this trivial tells me just how much i am:

useless

public procedure set_sendheader_useragent_msie() 
	set_sendheader("User-Agent","Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0)") 
end procedure 

{"User-Agent",": ","Opera/5.02 (Windows 98; U) [en]"}, --</joke> pick your own  

21. Re: The Archive and version compatability

• Posted by Kat Mar 28, 2009
• 914 views

Except declining my experience says my experience is worthless too. I have never run into a site where sending the hostname in the refer tag caused the webpage to not arrive. I would not have gone to the trouble of coding the default, if it wasn't needed, and i wouldn't have coded it if i had run into a site where it was a problem.

The rfc says the refer shouldn't be sent if the originating site isn't known, like if the user typed the url in from the keybd having never been to the host domain. It doesn't suggest anything more than that. And the entire document is a suggestion, it's a request for comments, it's not a strict standard, and there's a number of non-standards in wide use, like MS's IE. But if you are browsing around a site, then the host is indeed the origin of the urls you are clicking on. Rfc2616 wording on refer is identical to rfc2068 in this regard. It does not prohibit sending the source site of the url if you know it, it suggests you do send it. It doesn't mandate it be sent at all, ever, but some sites demand it, so i coded it.

useless