Make loop bounds consistent.

Is this what you mean?

Somehow, yes but not entirely. That solves the issue directly when reading the loops but I am more generally less at ease when we transition between counts and indexes. Translating an index to a count is nice but it still means we computed an index first, which is what confuses me in places.

So shall I calculate a list of counts rather than indices then?

I don't know... I am just worried that we are making code that is hard to read, follow and debug.

Let me read through the entire file and see whether I can grasp at a glance what is going now. That's all I am after in the end.

added 1 commit

• 135f0133 - Some minor style improvements. Bring the variable declarations closer to actual use.

Compare with previous version

Ok. I think I am happy with this version. I have made some additional stylistic changes. Let me know if you are happy with them.

Also, could you document what icount and icount_align are on lines 615 and 812?

added 2 commits

• 6a15034a - Commented definition of icount and icount_align.
• 7fb07a30 - Formatting.

Compare with previous version

Yeah, those stylistic changes make sense. I have documented icount and icount_align.

added 6 commits

• 7fb07a30...c1f773e7 - 5 commits from branch master
• bdf933d7 - Merge branch 'master' into fix_vec_names

Compare with previous version

BTW, there is no need to merge master back into every merge request unless the master has changes that git can't handle. That simplifies the reading of the logs.

Thanks for the changes! I'll accept this.

merged

mentioned in commit d9d199ee

Ah right, I thought it would fail if there were conflicts?

Can you tell if git can't handle it?

Yes, it will say the merge has conflicts at the top.

Ah okay, thanks. I will remember for future merge requests.

Conflicts should only appear if you are editing the same portion of code within the same files.