Error in compilation of bind-exit for JVM backend
Joseph Donaldson <[email protected]> Sat, 10 May 2025 18:21:50 +0000 (UTC)
| Newsgroups | gmane.lisp.scheme.bigloo |
|---|---|
| Message-ID | <[email protected]> |
------=_Part_3765205_1952732428.1746901310010 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Hello, Manuel, I just noted an error in the compilation of bind-exit for the jvm backend. = The error can be seen in the recette for the Pthread library. With some dig= ging, I was able to undercover that the problem is introduced when the retu= rn-goto optimization is enabled. When disabled, the recette runs successful= ly. Digging a little further, I narrowed the problem to the mark-set-exit! = procedure in comptime/Integrate/a.scm. If I comment out the second=C2=A0 fo= rceG? assignment, the pthread recette also runs successfully, even with the= return-goto optimizatino in place. Given your comment earlier in the proce= dure and the fact I am not completely sure I understand what is going on he= re, I have not submitted this change as a fix. Is this a reasonable modific= ation to fix the problem we are seeing in the pthread recette?=C2=A0 Also, I just submitted a pull request for changes to the Bigloo C runtime t= o accommodate C23 and GCC 15 compatibility. See https://github.com/manuel-s= errano/bigloo/pull/133. When you have time, please take a look.=C2=A0 Thanks,Joseph Donaldson=20 ------=_Part_3765205_1952732428.1746901310010 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable <html><head></head><body><div class=3D"yahoo-style-wrap" style=3D"font-fami= ly:Helvetica Neue, Helvetica, Arial, sans-serif;font-size:13px;"><div dir= =3D"ltr" data-setdir=3D"false">Hello, Manuel,</div><div dir=3D"ltr" data-se= tdir=3D"false"><br></div><div dir=3D"ltr" data-setdir=3D"false">I just note= d an error in the compilation of bind-exit for the jvm backend. The error c= an be seen in the recette for the Pthread library. With some digging, I was= able to undercover that the problem is introduced when the return-goto opt= imization is enabled. When disabled, the recette runs successfully. Digging= a little further, I narrowed the problem to the mark-set-exit! procedure i= n comptime/Integrate/a.scm. If I comment out the second forceG? assig= nment, the pthread recette also runs successfully, even with the return-got= o optimizatino in place. Given your comment earlier in the procedure and th= e fact I am not completely sure I understand what is going on here, I have = not submitted this change as a fix. Is this a reasonable modification to fi= x the problem we are seeing in the pthread recette? </div><div dir=3D"= ltr" data-setdir=3D"false"><br></div><div dir=3D"ltr" data-setdir=3D"false"= ><br></div><div dir=3D"ltr" data-setdir=3D"false">Also, I just submitted a = pull request for changes to the Bigloo C runtime to accommodate C23 and GCC= 15 compatibility. See <a href=3D"https://github.com/manuel-serrano/bigloo/= pull/133" rel=3D"nofollow" target=3D"_blank">https://github.com/manuel-serr= ano/bigloo/pull/133</a>. When you have time, please take a look. </div= ><div dir=3D"ltr" data-setdir=3D"false"><br></div><div dir=3D"ltr" data-set= dir=3D"false"><br></div><div dir=3D"ltr" data-setdir=3D"false">Thanks,</div= ><div dir=3D"ltr" data-setdir=3D"false">Joseph Donaldson <br></div></div></= body></html> ------=_Part_3765205_1952732428.1746901310010--