| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Quite a lot of them aren't in the same syntax/style, and it's quite bad
to look at, this should fix them all without causing issues.
|
|
I know, commas aren't needed, however, going in and out of using commas
and not using them also looks bad, so generally I try to always use
them, with exceptions.
|
|
This was supposed to always be there, however I somehow forgot to add
the winAlert() function, and forgot about it, causing some verbose
errors instead, not useful!
|
|
This should allow you to install packages that have dependencies,
however maybe not with the best UI/UX experience, as currently there's
only an English localization, and we also install dependencies even if
the dependency is already installed.
|
|
|
|
Everything seems to be working just fine, and I can't find any problems,
so I'll likely merge this soon.
|
|
Instead of repeating code and everything, get() now only returns a
boolean, based on the arg inputted into get(mod), it returns true if the
mod is enabled and false if disabled.
If the mod isn't found inside the enabledmods.json it is assumed that
it's enabled, as Northstar doesn't update this unless you actually
disable the mod first, and the mod won't be found in there.
Then I used get() to simplify list() a whole lot, and adding that
functionality of using the manifest for a version number if one couldn't
be found in the mod.json.
|
|
|
|
Albeit mods.list() doesn't actually search the manifest file for a
version number, and I'll implement that later.
|
|
|
|
Essentially just validates the config file and then prompts you about
it, it allows you to reset it directly or just to exit and let yourself
fix it. And because the error message appears directly in the renderer
we have access to navigator.language, and can therefore still localize
the string. However! We can't actually care if the user has disabled
auto detection of their language, since... y'know, the config file where
that's stored isn't able to be read properly.
And so I added an argument to lang(), which allows you to force it to
use a specific language if that language is available, if not it
defaults back to English.
|
|
Because viper.json doesn't exist it also means settings.gamepath is
null, which triggers the gamepathlost message.
|
|
I don't really understand why or how anything broke, all I know is this
fixes it, and I don't know how to feel about this madness.
|
|
|
|
When Viper starts up it'll check to make sure the gamepath still exists,
and throws errors if not, it also redirects you to the first page (the
one where you can set the gamepath), and gives you an informative error.
This could happen because the user unmounted the drive the gamepath is
on, or it could happen if the user moved their game location.
|
|
This allows someone to have their system in any language, and then have
Viper in a separate language. This is also useful for testing.
|
|
I forgot about this when merging the VDF fix
|
|
* added debug messages
* more debug logging
* even more debug logging
* removed debugging and fixed VDF problem maybe?
Apparently the libraryfolders.vdf doesn't always come with the same
values, sometimes it comes with the contentstatsid variable and other
times not, we assumed it always was there, and never checked for it,
this caused problems if it wasn't there. This should fix that...
|
|
If the nsargs are edited by a third-party program or anything that isn't
Viper, the next time you launch Viper it'll reset the nsargs back to
what it was when you last opened it.
|
|
When #87 was merged an unnoticed problem slipped in where if Viper was
running in a version that did not support auto-updates we would never
auto-update Northstar. Now that's fixed, and the nsupdate option is
completely implemented.
I also took the liberty of removing the auto-update option in the
settings page if Viper can't auto-update anyway, as it's useless if
that's not possible.
|
|
Albeit with some caveats, due to the way electron-updater is setup, if
you're not on a version of Viper which supports auto-updates the event
which runs the NS Updater code will never be run, and hence it'll never
auto-update. This was also an issue before, but I only noticed it now.
|
|
|
|
It actually saves settings, loads them properly and everything...
|
|
|
|
Albeit only frontend functionality, it doesn't actually save your
settings, it simply loads them, and Settings.get(), allows you to
convert them to a format that can be used to save settings.
|
|
Usually utils.js would default to simply unzipping whatever was uploaded
if it wasn't a folder, now we check if it has "zip" as the extension,
case insensitive. If not simply report back the file wasn't a mod.
As opposed to the obscure error you got before this.
|
|
|
|
incomplete update process
|
|
|
|
I have no clue why this is only a Windows error, but it is, which is
also why it slipped through into the 1.2.0 release.
|
|
The failed to install toast now pops up when it actually should.
|
|
The modpath variable is now refreshed when you change the game path.
Which stops the problem altogether.
|
|
It'll notify of incorrectly formatted mods, it'll also notify on errors
and on success.
|
|
|
|
|
|
We now directly check the manifest file info as well therefore in all
cases even if the folder name for the mod is completely obscure, we
should now finally still get an installed status, both on removal and
install.
|
|
If a developer for some reason decides to put their mod inside mods/
instead of putting it inside a folder in mods/, it used to fail the
install, now that's fixed.
However since there's no folder to directly copy the name from we take
it from the manifest.json provided by Thunderstore.
It of course also has the proper status in the browser UI
(update/reinstall etc)
|
|
|
|
A setInterval() will do the job, it's not performant, and I'll have a
look at improving this further down the line to something better.
|
|
I also fixed the archive not being extracted properly for some reason,
the setTimeout() seems to solve it, no clue if it needs to be adjusted
in the future or replaced with something that works 100% of the time,
I'm not even quite sure why it happens in the first place.
|
|
If a mod had a file that was bigger than 512mb inside it, it'd error
out, simply switching the copy-dir module to recursive-copy fixes this.
|
|
When a new mod has been installed through the browser UI it's button is
changed from "Install" to "Re-Install" instantly. We also now disable
modding related buttons when updating NS, or when installing mods. This
should prevent issues.
|
|
This should work for all mods, assuming that all mods come packages the
same, aka, in a Zip, with mods/<mod> in it. Which from what I know they
do.
|
|
This also fixes mods not showing up after first install of NS
|
|
When changing the game path by clicking the button you shouldn't be told
it can't find the game and you've to select one manually, as you know
that already. More importantly, if it could be found automatically it'll
just not do anything.
With this change I also changed the "gui.setpath" string to be more
logical, and to make it clear what it does.
|
|
* Edge case
* Alert user only when automatic detection failed
* extra info on not found message
* support for [Free/Open]BSD
Co-authored-by: 0neGal <mail@0negal.com>
|
|
|
|
I moved requests.js into extras and made the function for finding the
potentional gamepath into it's own module. I also made the exec() called
Promise based.
|
|
|
