refactor get-timing & glm

[dominikbach]

pspm_glm re-sorts conditions across multiple sessions. The same functionality is required for pspm_extract_segments. Hence, I moved this functionality to pspm_get_timing from where it is accessible to both functions.

[teddychao]

• pspm_get_timing
  • Line 227 -- 235
    • The function now has a new variable allnames to record the name from all sessions. The logic is sensible to me.
    • I tested how line 228 to line 234 works. I find the variable allnames can be updated when there are new values from in.names{iCond}. The updated allnames becomes an array, such as {'a1','a2'}. I expect in.names{iCond} to be always a simple string so that the strcmpi can be executed to see if allnames has included in.names{iCond}.
  • Line 298 -- 315
    • It seems to be checking if a condition (the session name is recorded in allnames{iCond} and is unique in it) and then obtain the tags from temptiming(iFile). I find the logic makes sense to me.
    • I think line 308 -- 310 seems to be a lower-level if clause under the condition of line 305. If my understand is correct I suggest to indent line 308--310 to a lower level.
• pspm_glm
  • Can I confirm the condition of line 398--415? It looks like lines 395--414 are all related to the condition of iSn. The current logic will make line 396 be executed at the same condition of line 409--414, but line 396 is separated written. Would it be easier to read if more line 396 to the level of line 409--414?

[dominikbach]

Thanks @teddychao for spotting these stylistic points - they are fixed now.