Glenn 'devalias' Grant

Crossposted: x.com/_devalias/st...

Consider evaluating identifier naming in a future version of the benchmark · Issue #2 · Ch3nYe/JsDeObsBench The following idea / note came from discussion on j4k0xb/webcrack#189: Afaik the evaluation metrics don't consider renamed identifiers so humanify would also have a pretty similar or the same score...

Also, interested if anyone knows of research or benchmarks for assessing/scoring the quality of variable/function/identifier names? Goal: to rank/benchmark restored names after reversing minified code.

Related: github.com/Ch3nYe/JsDeO...

Are there any similar benchmarks / leaderboards for JS unbundling / unminification? · Issue #3 · Ch3nYe/JsDeObsBench The following idea / question came from discussion on j4k0xb/webcrack#189: And for comparing wakaru it would be a better idea to use a dataset for minifiers, transpilers and bundlers. Yeah, that's ...

Anyone know of a good JavaScript Unminify benchmark by chance?

(Or want to build one)

I’m thinking of a similar vibe to JSDeObsBench: jsdeobf.github.io

An issue I opened for it previously: github.com/Ch3nYe/JsDeO...

Glenn 'devalias' Grant on X: "That feel when you ignore your 'notify on post' notifications for like 2 weeks, and then you finally go to procrastinate by catching up on them, and scroll and scroll, and then instead of actually reading them you start to wonder how you can count how many need triaging..." / X That feel when you ignore your 'notify on post' notifications for like 2 weeks, and then you finally go to procrastinate by catching up on them, and scroll and scroll, and then instead of actually reading them you start to wonder how you can count how many need triaging...

Crossposted:

x.com/_devalias/st...

www.linkedin.com/posts/glenn-...

Today's deepdive was probably also the most productive one yet for finding the bits and pieces I need to pull together a user script / Chrome extension to help make this 'keep up to date' / 'triage' process much less painful / time consuming...

But that's a story for another day

...but by that stage you've run out of time in the day to actually look through them all... 😅

At least I can quantify how much effort it will take now; which should help remove the 'do I have enough time for this' paralysis and maybe actually process them more often.

...and then with all of that, you build up some little helper functions and gadgets to stitch the data together, then a few more to process that hydrated array and count the tweets per day/user/etc, and then finally end up knowing that you have ~560 tweets to look through...

...but then you remember about React fibres and start looking through those to see if you can find where the full list of notification tweets are, and you find it but it's just a list of IDs, so you crawl deeper into the rabbit hole until you get a reference to the datastore...

...so you accidentally find yourself in the familiar rabbit hole of skimming through the DOM looking for relevant bits to build selectors off, and then remembering that the timeline only renders a slice of the list to the DOM so you can't just count rendered elements...

That feel when you ignore your 'notify on post' notifications for like 2 weeks, and then you finally go to procrastinate by catching up on them, and scroll and scroll, and then instead of actually reading them you start to wonder how you can count how many need triaging...

Crossposted:

x.com/_devalias/st...

www.linkedin.com/posts/glenn-...

For now the code/extension is unpublished, but eventually will be made open source and probably published to the Chrome extension store as well

48 windows with 121 tab groups and a total of 954 open tabs is fine.. right?

(Note: that 'open' is a bit nuanced, as I have Chrome features enabled that unload/freeze/etc closed tab groups / tabs that haven't been used recently, etc; without necessarily 'closing' those tabs)

This afternoon I decided to hack in a new feature to tell me how many Chrome windows, tab groups, and tabs I have open; so now I no longer need to guess/count how big of a problem it is!

A few weeks back I casually hacked together a PoC vibe coded chrome extension to help work around some of the annoyances I find with Chrome's Tab Group UX currently; and every now and then I've been making little tweaks to it to make it a bit nicer/more useful.

And similar to my original gist, to better direct people to the appropriate details in a slightly less scattered/chaotic way than in the past:

Serum 1 Preset File Format (.fxp):

gist.github.com/0xdevalias/5...

Serum 2 Preset File Format (.SerumPreset):

gist.github.com/0xdevalias/5...

I also added 2 new sections to my new gist, summarising the Serum 1 preset (.fxp) file format notes:

gist.github.com/0xdevalias/1...

And the Serum 2 preset (.SerumPreset) file format notes:

gist.github.com/0xdevalias/1...

Support for legacy Serum v1.x `.fxp` preset format · Issue #1 · KennethWussmann/serum-preset-packager I know the README explicitly states that it isn't supported: The legacy .fxp is NOT supported. But I figured since sooner or later someone is probably going to ask about it anyway; I'd preemptively...

The Serum v1 .fxp file format hasn't been fully figured out yet, but I summarised links to my old notes/tools/progress on this in the following issue, in case it helps anyone else make further progress on it:

github.com/KennethWussm...

Speaking of the discussions in the comments on my gist, KennethWussmann was making good progress on the Serum 2 preset file format: gist.github.com/0xdevalias/5...

And has since figured it out: gist.github.com/0xdevalias/5...

And published a tool to decode it to JSON: github.com/KennethWussm...

Glenn 'devalias' Grant on X: "Speaking of old projects of mine; there's been some interesting discussion recently on one of my old gists that in part is about reverse engineering / figuring out the Serum VST preset file format." / X Speaking of old projects of mine; there's been some interesting discussion recently on one of my old gists that in part is about reverse engineering / figuring out the Serum VST preset file format.

Crossposted: x.com/_devalias/st...

If reverse engineering / figuring out the Serum VST preset format is something you have any interest / insight in, I'd love to collaborate on it with anyone who's interested.

(I find it hard to find the time to push forward all of my projects as quickly as I would like to)

Which I then summarised/mused over to fill in the context as best I could remember it in this comment back on the main gist:

gist.github.com/0xdevalias/5...

Reverse Engineering Serum Preset Format - Supplementary Notes Reverse Engineering Serum Preset Format - Supplementary Notes - 1-reverse-engineering-serum-preset-format-supplementary-notes.md

This also inspired me to dig out some of my old notes and code (that I had forgotten where they were it had been that long.. 😅), and made a new gist to upload those to:

gist.github.com/0xdevalias/1...

The main gist is here:

gist.github.com/0xdevalias/5...

But check the comments for recent discussions and insights, for both Serum v1 and v2 preset file formats:

gist.github.com/0xdevalias/5...

Speaking of old projects of mine; there's been some interesting discussion recently on one of my old gists that in part is about reverse engineering / figuring out the Serum VST preset file format.

Made a few more tweaks to my PoC web app build diff minimiser script the other day.

It now can de-noise my 33,399 line sample diff down to just 3,991 lines; much easier to review/see the salient changes.

See the posts in this thread (on Twitter / X) for more details:

x.com/_devalias/st...

Glenn 'devalias' Grant on X: "Does anyone know if there is anything like https://t.co/HWyFB1Nx9a / https://t.co/wvbBKZFfzK but for the @npmjs / JavaScript ecosystem?" / X Does anyone know if there is anything like https://t.co/HWyFB1Nx9a / https://t.co/wvbBKZFfzK but for the @npmjs / JavaScript ecosystem?

Crossposted: x.com/_devalias/st...

This was the deep dive rabbithole of things I found when I was last looking into this sort of thing: gist.github.com/0xdevalias/3...

pypi-data pypi-data has 348 repositories available. Follow their code on GitHub.

Does anyone know if there is anything like github.com/pypi-data / py-code.org but for the @npmjs.bsky.social / JavaScript ecosystem?

"A Slice Of" Modern Program Analysis - Kyle Martin

Kyle's talk at Insomni'Hack is live! youtu.be/I0PoE0IdtmE?...

Check it out if you're interested in a slice of modern program analysis and try the latest version of Tanto as well, in the plugin manager or at github.com/Vector35/tanto