I'm not sure how you folks missed this but ALL diode lasers have the common problem that they will appear as a line. The diode structure itself will result in the optical resonant cavity having a higher divergence in one plane. Some very expensive diodes under the name circulaser had reduced that factor, but it is dificult to eliminate.http://www.newport.com/Tutorial-Laser-D ... ntent.aspx
However, divergence of the light emitting from a laser diode is very pronounced with full width half maximum (FWHM) angles of up to 40 degrees in the perpendicular axis (θ⊥) and 10 degrees in the parallel axis (θ||). This divergence results in a rapidly expanding elliptical cone. Gain-guided laser diodes tend to have greater differences between the two angles than index-guided laser diodes. Figure 4 illustrates beam divergence in the theta parallel and theta perpendicular axes.
Figure 4—Schematic representation of an index guided device and its diverging output beam profile.
Thus, the key takeaway is that the only true circular laser beams either come from gas lasers (argon, krypton, HeNe, HeCd, etc..) or, from diode pumped solid state (DPSS) lasers because they have larger optical cavities and thus limit the divergence in all planes resulting in a circular beam. While the effect is minimized at the focal point of the lense, it's should be obvious it's impossible for perfect focus if one axis is ~40 degrees divergent and the other is ~10 degrees.
Also, this would mean that blue diode lasers such as those from Blueray drives exhibit the same divergence as the typical red diode laser. Even the infrared diode laser does this in a DPSS, but as stated, the DPSS also has an external cavity and thus you never directly see the divergence from the diode.