On Tue, 17 May 2022, Sven Schreiber wrote: > Hi, > > with a Windows snapshot from April I found that the sample script of the > fsboost contributed package produces an error. This was weird because > the sample script of a package is always checked before approving it. So > I updated to the latest Windows snapshot from May 4th, but now gretl > crashes completely when running the same sample script. > > I suspect there's nothing wrong with fsboost, and instead the snapshots > after the 2022a release perhaps include some serious bugs leading to > crashes - I'm also thinking of the other recent problem on the mailing > list about saving session icons. I have no time to debug this properly right now, but I found that the crash occurs here on linux if the "ridge" package (which is a dependency) is not installed. After intalling that, the sample script of fsboost runs just fine.

Am 17.05.22 um 17:26 schrieb Sven Schreiber:

> > But it seems to confirm that the problem is not with fsboost. Hi folks, I already told Sven so but here it's just for the public list: Unfortunately, I won't have time to look at it before Thursday.

Am 17.05.2022 um 22:43 schrieb Sven Schreiber: > Am 17.05.2022 um 17:29 schrieb Artur T.: >> Am 17.05.22 um 17:26 schrieb Sven Schreiber: > >>> >>> But it seems to confirm that the problem is not with fsboost. >> >> >> Hi folks, >> >> I already told Sven so but here it's just for the public list: >> Unfortunately, I won't have time to look at it before Thursday. > > No need, probably. I checked with the 2022a release and everything runs > as expected. Again, some problem(s) seem to have been introduced in > gretl git since then. With yesterday's snapshot on Windows the sample script from fsboost does not produce a crash anymore. However, there's again the error that has been diagnosed as gretl now being stricter about the cnameset() function. That is, the fsboost package in at least one place uses cnameset() with arguments whose lengths do not match.

On Thu, May 26, 2022 at 5:27 AM Artur T. <atecon(a)posteo.de&gt; wrote: > > Am 19.05.22 um 18:03 schrieb Sven Schreiber: > > With yesterday's snapshot on Windows the sample script from fsboost does > > not produce a crash anymore. However, there's again the error that has > > been diagnosed as gretl now being stricter about the cnameset() > > function. That is, the fsboost package in at least one place uses > > cnameset() with arguments whose lengths do not match. > > Hi folks, > > thanks for the report. There was a mix of two issues: > > 1. A recent change in the cnameset() function requires that the length > of the string array matches the number of columns of the matrix. This > was not ensured (due to a bug) before. > > 2. Version 0.2 actually had a bug when calling the > plot_coefficient_paths() function. The function did not (always) > select > the actual final set of features before plotting. > > I just uploaded a new version (v0.3) which needs to get moderated and > which fixes the bug. Thanks, Artur. The ridge and Threshold_Panel packages have the same issue with cnameset (or rnameset in the second case).