hfe: Process opcodes when writing #825
Conversation
|
Also what's the test case? Is it particularly hard-sector index opcodes, or are other opcodes also tested? |
The start of a write corrupting OP_Bitrate applies to v3. That's hard to reliably trigger, but disasterous. Writing over OP_Bitrate in other cases is asking for confusion, but could probably be done a bunch without being noticed. OP_Index is important for v4, as otherwise writing to the last sector would break the track. Overwriting OP_Index could impact v3, but only sofar as you corrupt the image so it no longer works in HxC. |
|
Thanks. I added some more review comments. Hopefully the ones regarding |
There was a problem hiding this comment.
The w[i++] stuff worked out well.
I've made the changes, but it is essentially untested (I ran FF_TestBed, but that doesn't use hfev3). I'm going on vacation and not bringing Goteks, so it'll probably be two weeks before I can do that. I can rebase my file-cache branch on top of this while on vacation though, so it is ready when I get back.
|
Looking really good. Two minor comments. This can all wait until after vacation as I'm not making a v3.x release in the immediate future. I also want to rework the DSK, IMG streaming write code a little bit anyway, as I realised that in getting you to check for 128 bytes available in the main loop, I at least theoretically broke the Data Field CRC check (which needs only 2 bytes). I may end up doing more and if so I may make it a PR for you to review. |
|
Thanks for iterating on this. I've squashed it all together and committed to |
No description provided.