web2lrf

[kovidgoyal]

Hopefully fixed dmg is at http://theory.caltech.edu/~kovid/libprs500-0.4.44.dmg

Let me know if it works, and I'll release .45

[ddavtian]

Kovid, were there any changes in Newsweek profile? The one that you posted here on 3/12 (before feeds2lrf was released) was perfect. I got a similar one in 0.4.42. But 43 and 44 create different output (many more sections, one article per section).

I'm not qualified to add new profiles, but love using the built-in ones.

Also, there are two WSJ profiles in 4.44.

David

[Rick C]

Kovid,

I saw the new windows release of libprs500 0.4.44 while at work and d'lded it so I could have a go at creating a feed with your new manual, and was excited to see that a recipe for google reader had been added.

When I came home and d'lded it to my PC, I first uninstalled the old version and installed the new one. The main executable simply shuts itself down, no errors, no nothing.

I have cleaned out the registry, re-downloaded and re-installed/re-booted several times-there seems to be a problem.

[kovidgoyal]

Yeah the new newsweek profile downloads all the articles from the current issue rather than the RSS feeds. Oops, looks like I forgot to remove the old wall street journal profile.

[kovidgoyal]

@rickC

wait a bit for .45 (.44 has a bug that makes it crash on some systems)

[Deputy-Dawg]

Kovid,
The fixed version seems to work for me now. And with just one typed entry I was able to create a recipe for "Wispers in the Logia" RSS. It was one of those which was a bit vexing before!

Yes, I also had to do some cut and paste from Safari to get it to work.

[Rick C]

Well I was able to get v.44 to run on my other computer-unfortunately the google feed does not seem to work in this version.

If it can be tweaked,it would sure go a long way towards eliminating the need to learn python and still get a custom rss feed in lbprs. It is a great idea,really.

[kovidgoyal]

The present version of the google reader recipe only downloads starred feeds from your google reader account.

[ddavtian]

Google reader recipe for me downloads all the feeds (even not the starred ones) but only the summary part that's visible within a reader.

[kovidgoyal]

Hmm, I'll have to look at the code for the google reader recipe more carefully, it isn't mine.

[Rick C]

Originally Posted by ddavtian

Google reader recipe for me downloads all the feeds (even not the starred ones) but only the summary part that's visible within a reader.

OK I can get the starred items then, but not any of the others-and yes, I just get the first couple of lines like you said ddavtian.

[Deputy-Dawg]

Kovid,
I just downloaded 4.46 and I think the Gremlins are working overtime...

When I try to use my custom news sources it will only launch the one in the #1 slot and the one in the #6 slot. Or so it seems (I say that because I have not had enough time to try all permutations). And which one is launched seems to be predicated in some complex (or perhaps random) way on which was launched prior to the instant selection.

Fortunately it isn't going to get to much in my hair as much as I am now using crontab to do my news gathering. But it does make it a tad difficult to test the recipes for inclusion in the custom source list.

Sigh!!!

[kovidgoyal]

Wow, this has not been a good week for me. I'll look into it. You can always test new recipes by savin them in a .py file and passing the path of the file to feeds2lrf

[Deputy-Dawg]

Yeah, I am not at all concerned. And we all have bad weeks. Mine wasn't exactly sterling either. Monday I get up at 4 in the morning to be at the local hemodialysis center and am not on the machine until 9:15 starting a 3.5 hour treatment. I'm suppose to be 'on' by 6:30, and that was one of the better days.

I am off to bed.

BTW there is no pattern. On a recent pass the only recipe that would open was the one in the #1 Slot. Weird1

[Deputy-Dawg]

Kovid,
When I got up this morning I rebooted my machine and now the error is either much better and/or much stranger than before. The lead custom source in my list is a version of the Jerusalem Post that I use for benchmarking. When I click on it it does in fact download the Jerusalem Post. If I now click on my last entry, which is Whispers in the Logia - a new source using the newer 'recipe' I am told I am downloading the 'Jerusalem Post' but in fact it does download the Whispers in the Logia. If I click on the hourglass I am showen that, indeed, librs500 v4.46 is processing the Jerusalem Post. If I now click on the New York Times things get very strange indeed. My user name and password are embedded in the profile but I none the less get the following error message:

ValueError: The The New York Times recipe needs a username and password.
Failed to perform job: Fetch news from Jerusalem Post
Detailed traceback:
Traceback (most recent call last):
File "libprs500/parallel.pyo", line 139, in run_job
File "libprs500/ebooks/lrf/feeds/convert_from.pyo", line 40, in main
File "libprs500/web/feeds/main.pyo", line 126, in run_recipe
File "libprs500/web/feeds/news.pyo", line 816, in __init__
File "libprs500/web/feeds/news.pyo", line 393, in __init__
ValueError: The The New York Times recipe needs a username and password.
Log:
Fetching feeds...

Note that line 1 says that the New York Times recipe needs a username and password and that the second line says it failed to perform job: Fetch news from Jerusalem Post.

I have attached a pdf file containing 8 screen shots documenting my stepwise operations and what librs500 is displaying. It may very well be that this is what was going on last night but my connection was so miserably slow that I canceled what I thought were multiple downloads of files I had no interest in.