With the same data as in the bug before, if I create the 'time' variable and try to show its 'time seris plot' or the time series plot for a linear trend estimated over the 'precio' variable, the trend appears to have little stairs.
Two graphs with the 'not so linear' trend
What would you say the plot of the time trend _should_
look like, for 5-day data? Do you think it would be
better to put imaginary "5-day time" on the x-axis rather
than calendar time? (Clearly, the line is not going to
be smooth if we use calendar time, as we now do.)
Based on your comment, I see it is a matter of interpretation.
It is not really a bug.
I did not notice that even the time series plot of the 'precio' variable was plotted against 'real time'.
I think that the user is already making a decission when he organizes the data as diary (5) or diary (7). If the user has 5 observations per week, but (for a reason I don't know) decides to consider the weekends as 'NA', he can do that by selecting diary(7). By contrary if the user has selected diary(5), I think he would like to see in the graphs observation 6 just after observation 5 with no spaces between them. So in response to your question: yes, I think it would be better to put imaginary "5-day time" on the x-axis, because it would be more consistent with the behaviour that a time series analyst is used to see with other periodicity data. But, I have never worked with diary data, so I may be missing something.
Given that it is not really a bug, I think you can change the status to closed, and If you consider interesting to have a wider consensus about this, please consider to share this question with the gretl-devel or gretl-users list.
Following Ignacio's suggestion and closing this.