adamlaurie / rompar Goto Github PK

For OS' that should conventionally be GUI only (such as improving Windows support: #17), the common cases should not require any command line usage

However, as implemented a project must always be open. Need to figure out some way to work around that, whether it be starting the workspace idle, or popping up a new project wizard if no arguments are given

Its possible to run on Windows, but requires knowing about things like pip. It sounds like there are some options to create simple MSI installers. Maybe we can get a volunteer to work on this? @diamondman might be able to provide some direction, but not sure if he has time he can devote right now

"Either use the NSIS tool (https://nsis.sourceforge.io/Main_Page) or MS has some tools for generating MSIs from a set of files and registry keys"

Right now if you give an invalid file name scary stack traces pop up. Review bad user input possibilities and make something friendlier appear

What should this flag do specifically? Does it do the right thing in the current code?

Related: #13

We need a way to visually compare multiple binaries against original images. I'm thinking:
-Start with a reference rompar project w/ an image and bit layout defined
-A way to load in a new binary, replacing existing bit definitions. Probably by using zorrom to convert to .txt, and then loading in the .txt to rompar
-A way to highlight suspicious bits

both say minimum

Hi Adam, John,
You remember I could not get rompar working last week at hardwear.io?
Actually it was not a problem within rompar but because of opencv.

In case rompar fails starting with an error such as

QObject::moveToThread: Current thread (0x2956650) is not the object's thread (0x2bd9990).
Cannot move to target thread (0x2956650)

https://stackoverflow.com/a/63839383 indicates it results from the duplication of libraries both on the host and the opencv-python distro.
To solve it, one can recompile opencv-python:
pip install --no-binary opencv-python opencv-python
Beware it takes quite a while, but it solves the problem :)

Currently rompar lacks an undo feature, and there are many actions that are rather annoying to manually undo from the UI, so a reload from a backup or text-file edit is required.

A quick and dirty solution I am implementing right now repurposes the "load from file" code and maintains a stack of previous states for undo that are generated using dump_grid_configuration.

I was wondering if such a solution would be welcomed by the project, or if there are reasons to not use this implementation for undo?

When starting up a new project (e.g. without the --load command) the edit mode icon that is shown is the "edit" mode (e.g. the two dots) however the actual mode in operation is the "grid" mode (e.g. where you define the rows/cols), which is misleading.

See below for what the startup looks like:

In this case the "grid" button should be shown, as it is the actual selected mode.

Does this run under Windows? If so create instructions and add to README

https://github.com/peterbjornx/pladecode

Some of the features are:

• Better support for poorly aligned images
• Better support for rows squished close together, blending nearby images

Consider maybe a pip package to ease installation

https://pypi.org/project/rompar/

This is woefully out of date and is before the Qt conversion

Hi,

I've use this wonderful tool, and was impressed by it. Whaoo !
But, I found a little bug : when you save a file with "S" function, the binary dump have some "0D" (CR) added.
This issue can be solved by changing line 685 to :

outfile= open(basename + '.dat%d.set%d' % (x, Saveset),'wb')

If you open file with 'w' only, I guess it's ASCII and nasty chars appears.

Sorry for not submit you a nice patch, I've never code in Python nor use GitHub ...

Would be happy to see more stuff from Aperture Labs !

Cheers

Phil

Adam's article is out of date vs new Qt GUI. A video would be ideal that goes through a few different ROM types and how to tune to decode them cleanly

There are some APIs to delete grid entries, but doesn't look like they are connected?