shawn

it knows based on several different variables. 1) if you're using the "use following variable for indicator of changes" then it will use that variable. you can populate that variable yourself with the variables feature by scraping the source site for the version number, hash or some other unique value for each version 2) if you're not using a change indicator yourself, then ketarin will analyze the HTTP headers to determine if the download is the same file. it can do this without downloading the file by looking at several different headers, including "Age", "Content-Length", "Date", "ETag", and "Last-Modified". These headers (or variations of them) can be returned on a HEAD request, so the complete file does not have to be downloaded in order to determine if it is the same file that was already downloaded. Unfortunately, some sites do not behave well or do not include these headers. In these cases especially (but really, in *all* cases) you should be scraping the site first for a version indicator. If you want to FORCE it to download then you need to use the force option (Ctrl+F5 / aka "force download"). This method will force the download to be attempted again, and re-download the file, even if Ketarin does not believe it is a new file.

thanks @floele!

note on passwords - if the password has a special character then quoting will help but in some cases (such as caret, pipe and doublequote) will still need to be escaped for it to work. Use a caret to escape most characters, and echo to test it: ECHO. This is the quoted password: "Princess^|1"

Based on those settings I suspect the problem is most likely the unquoted password or LSA access. Passwords tend to have non-alphanumeric characters, and if the password has a caret (^), ampersand (&), parenthesis ("(" or ")"), space, doublequote ("), or one of several other characters then it'll fail to authenticate because those characters are being treated differently from the command line. You should be able to check the errorlevel to determine if that's that case: swithmail.exe ... IF %errorlevel%==1 ECHO.Something is wrong with sending email. If it's really a Gmail account then the other likely issue is that you haven't yet enabled LSA/"less secure apps" for that account. Verify that this is enabled and try again.

Fantastic.

Yes. Use the option "Ignore file information (do not redownload if file is missing)". This seems to be a popular topic right now.

ah - filehippo. hash checking is imposed on all filehippo apps, and unfortunately, their site sometimes doesn't work or consistently allow for content scraping. currently, the MD5 hash is on the "Technical" page here. while the MD5 does match that of the file, it's most likely a matter of scrape failure. @floele - it looks like there might be a problem with FH again

It should also be noted that the nomenclature for these specific xml files are called "applications". and there's a reason for that. Ketarin also has template support which allows future apps from a given provider (such as sourceforge, wordpress.org, or something similar) to be able to have most options prefilled and allow drop-down population for other options. Check it out.

you can also use a dotted path ("..\log.txt") - which is what I do for my log.

yeah, i'm afraid email notifications from the forum are rather hit and miss. mostly miss. they usually come from 136.243.175.196 which is hosted via Hetzner Online (aka Hos-Gun) which has a *lot* of spammers hosted as well. Not on that IP, but on the same provider, which tends to mean the entire provider gets blacklisted by many mail services. you might have better luck by adding that IP and noreply @ canneverbe.com to your whitelist.

the real difference is context. "post-update" is still within the download/update process, so it's best used for things like purging archival copies, fluff information or passing off the download for secondary processing (such as torrents to aria2). you can still use flags at this point to indicate whether it's succeeded or failed. i think the error you're seeing is really more about the process termination. i'm not familiar with the application you're using, but i've had great success with pskill: pskill -t Ditto.exe This kills all processes that match that name *and* all child threads. It also doesn't "return" to continue processing until it either succeeds or fails. i suspect that the other one may continue processing before completely killing the process, which means the next step won't work (copying over the new file). The first time you run it you'll need to "allow" it and accept the license agreement, which stores a registry flag so you won't be prompted again in the future.

On each individual affected application you can check the box on the Advanced tab that reads "Ignore file information (do not re-download if file is missing)" to get the same effect. The version number is stored within Ketarin (as long as it's been previously downloaded at least once) so you don't need to keep the archive of apps locally. I use this for large blobs like zipped ISOs and torrents that are ultimately stored in different paths than where the Ketarin app cache is located. I'm happy to provide an example if you would like. I think most of your issues are related to that checkbox. As for the exit codes - I agree, there should be a pre-update exit code to effect an "ok/current" status. @floele can you please implement this?

The online database can be found here, or by clicking the chevron next to "Add new application" and selecting "import from online database"

And is the hash-value drop-down set to "(None)" or MD5?

do you have a variable or other value in the Advanced, "Compare with hash value" setting?

Updated

@necrox - can you either share a link to the specific app in the repository or share your XML? this will help us troubleshoot. thanks!

it looks like you might be using "check for updates only, do not download". Please post the XML for the 7z or Firefox so I can verify this.

this doesn't redirect to mediafire - so probably not. but it does consistently download the file without touching mediafire, which is my preferred behavior.

There are many samples in the "Templates" section for various sites. There are also over 3000 samples in the Applications repository - and you can search it based on keyword (like sourceforge) to see how others are working with specific sites. The problem on most sites is that while they are a file distribution network of sorts, they don't necessarily have the most current file in a very consistent naming pattern, format and style, so selecting it from the page so you get the correct file is more challenging. Here's a sample image I collected for an article I haven't had time to write yet: On Sourceforge, for the FileZilla Client, they're claiming the "latest version" is 3.17.0.1 - which is over a year old, 20 versions behind, and has at least 8 different security vulnerabilities that have been patched in the current release. This means that you can't trust the sourceforge "latest version" URL, and have to parse the directory structure that the developer created for their releases. Yes, it's a pain, but since it's not reliable any other way, it's the way it has to be done. I prefer to use the original sites anyway. Yes, it requires more labor initially, but it's the only way to know for sure. Some sites, including sourceforge, have been blacklisted by developers since they started click-wrapping their installers with adware packages. It's not just evil, it violates their terms of use and they'd rather not have a distribution network infecting their users that make it look like the developer is at fault, when it's clearly the distributor.

It really depends on how you have it configured. you'll need to assign it to a variable that exists - which means one that you've created within the [Variables] button on the Application tab. If the variable "version" doesn't exist yet, then you'll need to make one and populate it by scalping a webpage where the version is displayed, either with a regular expression or start/end behavior. If you share the link where you're getting the download for the problem app, I'll demonstrate how to build a version variable.

You need to make sure that you're using an "indicator for changes". You'll find this on the second tab (Advanced settings) of the main app window. Populate the field "use the following variable as indicator for changes" with the variable (such as "version") that you're monitoring to determine if an app update actually occurred. Without this field populated, Ketarin has to perform a best-guess effort to determine if a file has changed. If the server doesn't support E-Tag, accurate timestamps or other common detail information about the download, then you will end up with Ketarin downloading the file *every* single time you check for updates, even if nothing has changed. If the standard HTTP methods of identifying changes are not being offered by the server, the only way Ketarin can know is to complete the download and check against the existing file. For small files this is a minor inconvenience. For OS ISOs and the like it can be a huge waste of resources.

hi, @floele. application search (within the forum) still isn't working.

tested and working. thanks, @floele!

thank you, @floele !