http_get_page2

[MikeC]

Re: http_get_page2

If I remembered what it was I would have given it when I posted...seems familiar when attempting some API/Windows work using Alpha to obtain info but cannot place the exact function.

The main point is that some Alpha functions halt the code and some don't.....and that the documentation usually does not specify which (although sys_shell() does inform us that your scripts will continue to run). UDFs, at least all that I have created, seem to run one at a time--using UDFs have cured an enormous amount of timing issues I have had in my work.

[G Gabriel]

Re: http_get_page2

TThe issue described above seems to indicate that at least some A5 functions will return control to the calling script before they have completed their task; i.e., all functions are not created equal!

Not sure how you arrived at that conclusion?

I have run across/mentioned this in the past as well....I remember an Alpha function that even was inconsistent in how it acted...sometimes would finish prior to the next and other times not.....I now assume that none but UDFs work in succession in order to make certain aspects of an application more bulletproof.

Never heard that dog bark before! Would like to see an example.

[MikeC]

Re: http_get_page2

I have run across/mentioned this in the past as well....I remember an Alpha function that even was inconsistent in how it acted...sometimes would finish prior to the next and other times not.....I now assume that none but UDFs work in succession in order to make certain aspects of an application more bulletproof.

[CALocklin]

Re: http_get_page2

Yes, that did the trick. I changed the script to have it call a UDF. The only thing the udf did was get the source html using http_get_page2(), and return it in a char variable. My script then waited for the function to do its work before beginning my parse routines.

That should be a point of general interest. It's something I've suspected for a long time but haven't bothered testing.

Apparently UDFs will wait until everything inside has completed before returning control to the calling script/function but A5 built-in functions do not necessarily operate the same way. The issue described above seems to indicate that at least some A5 functions will return control to the calling script before they have completed their task; i.e., all functions are not created equal!

It would be interesting to know which built-in functions act this way. It may be the reason for all the weird, totally random crashes in my most complex application.

[Tom Cone Jr]

Re: http_get_page2

Thanks all.

Further digging here indicates that maybe this is a timing issue. I'm trying to retrieve the source html from the specified web page and then parse it. This is happening in a script since I'm trying to automate the sequence. I'm beginning to think maybe, just maybe, what's happening is that my script is beginning to parse the source text before it's all been downloaded. Since my script looks for markers in the source text that are near the bottom of the web page its returning my "not found" error trap, probably well before the source itself has been completely downloaded. Is this possible?

What's odd is that week before last it ran cleanly.

Anyway, I'm going to move the "get source" routine to a separate UDF to isolate it from the script. Will let you know what I find. Thanks again.

Later on ...

Yes, that did the trick. I changed the script to have it call a UDF. The only thing the udf did was get the source html using http_get_page2(), and return it in a char variable. My script then waited for the function to do its work before beginning my parse routines.

[Garry Flanigan]

Re: http_get_page2

Tom,
In my earlier post I should have written http_post_page2() not http_get_page2().

So have you tried http_post_page2() with

"include_headers" = .t.

?

[MikeC]

Re: http_get_page2

Garry,
HTTP_GET_PAGE() obtains the headers by default and does not obtain the "Includes" that are missing.....using the header argument in the HTTP_GET_PAGE2() function actually delivers a bit less information....

This sure seems to be something that will be wanted by possibly many and hopefully a way is found to get around the apparent limitation that is imposed by some websites.

[Garry Flanigan]

Re: http_get_page2

Tom,
Have you tried http_get_page2() with

"include_headers" = .t.

?

The default is .f.

[G Gabriel]

Re: http_get_page2

Tom:
Yes the page, like many others, almost all others, are built dynamically but that has nothing to do with the page info. That's static until the web designer changes it. The scripts that trigger those other parts of the page, those scripts are part of the page info and you should be able to see them.

These functions evidently aren't doing the job. I haven't tried others but I am not sure why does the client want to copy the page info? For what use?

[Tom Cone Jr]

Re: http_get_page2

Well, bad news. Today the http_get() function can't retrieve the body of the webpage either. This particular web site is a report that apparently is getting "built" dynamically. It seems the Alpha http_get() function is hitting the site and returning what it finds before the page is fully constructed. Have no solution for this other than telling the user to open the web page and manually copy down the data that's of importance. Bummer.

[G Gabriel]

Re: http_get_page2

Just to throw another monkey wrench in the mix:
I knew this issue was brought up before but couldn't quite remember all the specifics and with a little search I came across this Thread.

I'd still report it as a bug though I expect Lenny to say: I told you so!

[MikeC]

Re: http_get_page2

deleted as incorrect....

[Tom Cone Jr]

Re: http_get_page2

Thanks to a tip from Dr. Martin Heller I'll be able to rework my code using

HTTP_GET()

instead of http_get_page2().

Thanks to all who responded.

-- tom

[Tom Cone Jr]

Re: http_get_page2

Darn, that's not good. It's the "includes" that I'm trying to get.

Any idea how I might go about it?

Thanks.

-- tom

[MikeC]

Re: http_get_page2

Tom,
I think you just provided yourself with the reason why you do not get the entire content....you are getting the entire page...just not any of the "Includes".



EDIT LATER:I even made my default browser IE as the help file indicates that at least some of the functions use IE explicitly (don't know if that is actually still true)....no avail.