WORKFLOW: Automatically test importing save files from old versions

[catloversg]

In #1790, we fixed a bug related to migration code. Currently, our CI workflows do not test importing save files from old versions. In risky PRs, d0sboots has to manually test it manually. It's best to implement a workflow doing it automatically for each PR.

My plan:

• Use an E2E testing framework (Selenium, Cypress, etc.) to simulate loading save files and check regression bugs (Recovery mode UI, exception alert popup, etc.)
• Generate testing save files from old versions. d0sboots has a collection of user-provided save files, but using them publicly in a testing workflow is risky due to potential licensing issues. It's best to create our own collection.

This plan is fairly low on my priority list, so I created this issue to make sure that I don't forget about it (or procrastinate too much).

[d0sboots]

I think our existing testing framework would be sufficient. We don't need to test end-to-end behavior like "does the screen look like this," it is sufficient to ensure that the game manages to toPage somewhere else without crashing.

Also, the savegames I have tickle specific loading bugs we used to have, although we can probably recreate some of those scenarios. The point being, artificially created saves don't cover the same breath of code (sadly).

[catloversg]

I think our existing testing framework would be sufficient.

Maybe. Some regression bugs may only be shown with special conditions (for example, when processing invalid AllGangs, the game only throws an error when switching to the Gang page or processing the TW tick). When I have time to deal with this issue, I'll check if Jest is good enough for us.

[AdamAndreatta]

Hi!, i was wondering if this issue is still currently open and if so would i be able to pick if up to work on it. :D much appreciated

[d0sboots]

It's definitely still open. (Anything marked open is still open, unless we've forgotten to close it, which does happen but we try to keep it current.)
It's a pretty ambitious undertaking though, so although I won't say you can't pick it up, I might encourage you to get your feet wet with something else first unless you really have your heart set on this. :)

[AdamAndreatta]

No worries i can find something to get my feet wet : D. Is there an issue you guys have thats relatively simple or that feel is dimple in the scope the project.

[d0sboots]

OK, I went looking for some easier ones.
#1640 is one, someone even took a stab at it but wandered off halfway through (not uncommon)
Another one is #1504, although only if you have an opinion on it, otherwise not.
#1134 also seems straightforward.

Unfortunately most of the outstanding bugs are complicated or very opinion-based XD

[AdamAndreatta]

Ill take on the last issue #1134 thanks so very much for finding a couple. Im not sure how the assignment process works or if you flag that ill be handling this issue.

[d0sboots]

No assignments just go for it. If you finish it it'll be marked closed when it gets submitted.

[AdamAndreatta]

Quick question after building the project it says the start target index html does not exist for electron do i need to adjust the path myself. (npm found the proper build instructions)

[AdamAndreatta]

I apologize for bothering you again, but I have one final question. I finished writing the api it returns all the exp values gained when passed a set of parameters I wasnt sure however how to go about editing the documentation to i need to create an interface for netscript or something of that manner.

…

On Fri, Feb 7, 2025 at 4:41 PM David Walker ***@***.***> wrote: No assignments just go for it. If you finish it it'll be marked closed when it gets submitted. — Reply to this email directly, view it on GitHub <#1793 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A5TXIAYWRHCHYGNKSWE6UF32OVHCXAVCNFSM6AAAAABSKT54F2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNBUGM4DEMRXG4> . You are receiving this because you commented.Message ID: ***@***.***>