On Sat, Aug 30, 2025 at 08:19:03PM +0300, Eli Zaretskii wrote: > > It reproduces Gavin findings on OpenBSD. To reproduce the failure of > > cygwin (and mingw on cygwin) I added a variant where the second library > > is in a different directory than the dlopened library, which is more > > like the actual situation of Texinfo. This is consistent with what Eli > > explained on how libraries are found in windows. > > Can you modify the tests to use the absolute file name of the dynamic > libraries? Then I believe there should be no problems for dlopen to > find the libraries.
I modified Gavin code to try both the absolute and the relative path for the dlopened library. But the problem is not for the dlopened library, the problem is with the second library, which the dlopened library is linked to. Actually, dlopen fails both for the absolute and the relative path when the library dlopened is not in the same directory than the library the dlopened library is linked against. When the two libraries are in the same directory, I do not report which of the dlopen, with relative or absolute succeeded. I could report if the dlopen with relative path failed if it is an interesting information. -- Pat
