Typo in `?gelqt` documentation.

[Jutho]

Description

The documentation of the blocked LQ decomposition methods ?gelqt contains a typo. The size of the work array should be mb * m instead of mb * n. If the latter value is used for matrices where m > n, all kinds of segmentation faults or memory corruption errors occur. These disappear completely when a work vector of length mb * m is used instead, independent of the value of mb.

Checklist

• [ ] I've included a minimal example to reproduce the issue
• [ ] I'd be willing to make a PR to solve this issue