Release Commercial apps now accepted in the repo.

[fusion_power]

If you heared the storys I heared about PayPal, you maybe wouldn't use the word "trust" in combination with PayPal.


How about the other Competitors? Google Checkout, and wasn't there also an Amazon - thingy- 'n stuff as additional options?

 

[milkshake]

1 step at a time plus it means the devs would need those accounts too as non of the funds flow by OP

 

[Jey123456]

As much as i love google. i never liked google checkout, (or more precisely, they never seem to like my credit card heh).


Last time i had to use it, it refused my address because according to googlemap it couldnt exist (im in between 2 postal code zone). Took a few hours of try and error with my credit card company on the phone. In the end, we had to fake my adress on the credit card so just so i could pay >_> .


Not to mention the fact that unlike paypal, you cant use your bank account directly to pay. I did hear a few bad story related to paypal, but in all the years i used it, i never had any issues heh (other than the fact that more and more places refuse paypal because of their tendency to refund before asking). but thats an unrelated issue.

 

[Mr Rob]

And at the end of the day, you'll never be able to find a payment option to satisfy everyone.

 

[Perky_Goth]

I did hear a few bad story related to paypal, but in all the years i used it, i never had any issues heh (other than the fact that more and more places refuse paypal because of their tendency to refund before asking). but thats an unrelated issue.

That's because it's not the buyer who gets *ed, it's the seller. You can easily buy something on ebay and tell paypal it never arrived and get your money back with no recourse.


Don't do it, though.

 

[Tempel]

milkshake, did you ever look into digest authentication for repo clients? If it looks feasible to you, I'll write it into the spec. Then Cloudef and I can get on implementing it.

 

[milkshake]

not as of yet no, but I will, don't know when though.

 

[Cloudef]

I'll probably add that to the rewrite then. I'm ATM pretty much busy with life, but I try to start soon. So you probably need to depend more on Tempel currently.


@Tempel


Btw, I made minor change to PND repo specs that links should be always url encoded so that client does not need to do any preprocessing.


There was slight breakage in few apps on pndstore and milkyhelper because of this.


Also, I think I was saying once I do the python bindings, but never got around doing it. Maybe the better design will help that as well.

 

[Tempel]

Thanks Cloudef. I had just assumed that would be true, but it's better to have it mentioned explicitly. Was this breakage related to the repobreaker test that was going on earlier? Are there any packages that currently cause such breakage?

 

[milkshake]

only the repobreaker but that app is impossible to function correctly due to the crazy chars in the package ID but I now do url encode where needed so what broke before is fixed.

 

[Cloudef]

Thanks Cloudef. I had just assumed that would be true, but it's better to have it mentioned explicitly. Was this breakage related to the repobreaker test that was going on earlier? Are there any packages that currently cause such breakage?

It was packages with spaces and characters that need to be escaped, such as repobreaker, (and before that The Slime Whatever pnd) and before even that long ago there was id with some weird character don't remember name anymore, but milkshake apparently renamed that.

 

[milkshake]

in order to make things like repobreaker work im giong to have to create a new primary key and instead of using app ids for that use a number that way repobreaker would still work.

 

[Cloudef]

Hmm, repobreaker worked fine after the urlencoding? At least on milkyhelper.

 

[milkshake]

yes but the icon loading and preview pics loading didnt work, and I couldnt get it to work even with proper url encoding do to the folder name of where the preview pics/icons are stored use the app id as the foldername.

 

[Cloudef]

ah, yeah that.

 

[Tempel]

You could just reject package IDs that use certain characters. The PXML spec just says that the "ID must not include characters which would be invalid in directory or file names", which is a little unspecific. If you enforce a specific character set for IDs, I don't think it would go amiss.

 

[WizardStan]

In that vein, an additional feature that would be useful is if, upon detecting these invalid IDs (or, in face, any invalid PXML data) there was a way to edit it on the server so as not to have to reupload the entire thing again. Probably just dreaming there, likely to take more work than it saves.

 

[Cloudef]

You could just reject package IDs that use certain characters. The PXML spec just says that the "ID must not include characters which would be invalid in directory or file names", which is a little unspecific. If you enforce a specific character set for IDs, I don't think it would go amiss.

Yeah, I already bit of mentioned at IRC that PXML spec is sort of idiotic in many ways (not only that it uses XML), but having ID's that just take about any character (minus fat illegal filenames), is not quite cool IMO. But yeah, got shot down.

 

[milkshake]

yeah which is why im going to have to do it assigning my own local id's for the db bu still store the package id etc

 

[milkshake]

yeah its not cool but i guess Ivanovic is correct as linux accepts those chars in a file/folder name then they should be allowed (due to appdata folder name defaulting to app id), just makes thinks arkward for online distribution.