News:

Yahoo Groups closing on Dec 14th 2019

Main Menu

New test version

Started by Wilson Logan, May 24, 2009, 01:08:58 PM

Previous topic - Next topic

tecno


Wilson Logan

On the news front...

The sticking messages are caused by Yahoo randomly serving summary pages of data. PGO requests 30 messages at a time in expanded mode. Yahoo sometimes serves a page of just Subject headings.

Its odd because the pages that get served incorrectly are random in order but reproducable i.e. messages 34, 56, 78 and 89 may give errors but they'll do that every time.

We're going to check to see that the URL we request is not malformed but I think it can;t be or we'd get none of the 30 messages and we will try to request the problem messages as one page if the URL turns out to be good.

Cheers,

Wilson.

tecno

Any progress last week? I´m getting a bit frustrated as I have so many groups to follow up.

Wilson Logan

OK, it appears the URL is good but we are getting a mix of formats within individual groups. The bad thing is that the second format that is currently causing the 'message missing' errors is actually causing the parser to crash. To get technical, the parser is written in C++ and uses a library called Boost. Within this library are all the regular expression commands. The one we use most is regex_split but this command has been deprecated (ie is no longer supported). Because it is no longer supported we can't go to Boost and ask why the HTML we have crashes the command because they'll say "duh, its no longer supported, use the new command". So we're substituting the regex_split commands for regex_token_iterator but it doesn't work just like regex_split used to and there is a bit of work involved.

I shall write a separate post on why the heck this is taking so long.

Cheers,

Wilson.


MyNameIsScott

In my work we are evaluating OCR and data population tools, and the ones that we are looking at are designed to read multiple variations on a central theme.

It sounds like the program logic isn't so much an issue as it is the actual coding.

Maybe one of these software houses has a module that could be used?

Copanion, SurePrep, ProSystemFX scan, just to name three.

Just a thought ....

Scott

mjnelson99

Wilson,
I tried all my active groups and none came in. The ones that had no new messages did correctly tell me that. All of those that had new messages would time out and program rebooted.

One of my groups is set in Yahoo, Asia and it also did not work. I can get those messages to come through as POP directly into Thuderbird.

Since I am using a non-Yahoo user name for group messages I could just save all groups into their own folders for now until PGOffline is fixed. I know Thunderbird can save them and I am sure other email programs such as OE and Windows Mail can do that also so it could be an option for us all temporarily.

mjnelson99

Downloaded and replaced the dll Wilson put on a different post. Don't have post # to give.

I could then download all groups successfully. It looked like PGOffline was crashing after each download had completed, though.

Re-registered using C:\Windows\System32\regsvr32 MSCOMCTL.OCX from command prompt windows run as Admin.

Was also having problems compacting.

Maybe Vista was lazy or maybe PGOffline. I can now compact. I did not do anything else on the computer while it was compacting, though.

Since all downloads are already brought in I cannot tell if program is still crashing after it completes. Will have new ones to pick up tomorrow so can tell more then.

Thanks for posting that dll. It did the trick along with new test program. Not sure if dll alone would have worked. I would have tested that if I had seen the dll first.
Mary

Wilson Logan

#22
Hi Mary,

There's an even newer DLL I want everyone to try.

I'm going to post a new message in a second with instructions.

http://www.personalgroupware.com/smf/index.php?topic=719.0

Cheers,

Wilson.

SMF spam blocked by CleanTalk