Re: Claude AI code audit of GNUstep core stack — 150 fixes, 12 perf optimizations, all availa ble for upstream
Patryk Laurent <[email protected]> Mon, 13 Apr 2026 08:37:26 -0700
| Newsgroups | gmane.comp.lib.gnustep.devel |
|---|---|
| Message-ID | <[email protected]> |
--Apple-Mail-CF580FF3-E711-409C-B050-25FBA2C40AA3 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable I would advise that each change (each fix, each optimization, etc) be submit= ted as a separate reviewable PR. Ideally, signed off by the contributor.=20= Furthermore in the case of each optimization a benchmark should probably be d= one. Thank you, Patryk -- Patryk Laurent, Ph.D San Diego, CA w: http://pakl.net/ > On Apr 13, 2026, at 07:40, R Frith-Macdonald <[email protected]= > wrote: >=20 > =EF=BB=BF >> On 13/04/2026 04:35, Todd White wrote: >> Hi GNUstep Team, >>=20 >> As an exercise to test out the latest Claude AI capabilities, we recently= completed a comprehensive, bottom-up code audit of the GNUstep core stack =E2= =80=94 all seven repositories =E2=80=94 covering libobjc2, libs-base, libs-c= orebase, libs-opal, libs-quartzcore, libs-gui, and libs-back. >=20 > Thanks for that, it looks like very worthwhile work. I can't really comme= nt on other packages, but I'd certainly like to incorporate a lot of that in= to GNUstep-base. In the areas of the base library I can see quite a mixture= of stuff, some of which is already known, other parts are quite new. For i= nstance, at one extreme having server verification off by default was a deci= sion from many years back, but one where a change was long overdue, while at= the other extreme the check for cross-thread autorelease pool draining deal= s with such a perverse situation (I find it hard to concieve of any way to a= ccidentally do that) that it's quite delightful to have a check/fix for it (= though classifying it as a critical issue seems silly). >=20 > In an ideal world, I'd want to see an individual patch for each issue. Th= e patch would contain the source code change, the corresponding testcases, a= nd a ChangeLog entry, but the cloned repos contain only the first of those. >=20 > Can the LLM take the testcases it has generated and rewrite them as portab= le (excluding objc2 and/or platform specific features or controlling them wi= th preprocessor macros) code as used by the gnustep regression testing frame= work, and add them to the regression tests in each package? >=20 > Thanks >=20 >=20 >=20 --Apple-Mail-CF580FF3-E711-409C-B050-25FBA2C40AA3 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable <html class=3D"apple-mail-supports-explicit-dark-mode"><head><meta http-equi= v=3D"content-type" content=3D"text/html; charset=3Dutf-8"></head><body dir=3D= "auto"><div dir=3D"ltr"><br></div><div dir=3D"ltr">I would advise that each c= hange (each fix, each optimization, etc) be submitted as a separate reviewab= le PR. Ideally, signed off by the contributor. </div><div dir=3D"= ltr"><br></div><div dir=3D"ltr">Furthermore in the case of each optimization= a benchmark should probably be done.</div><div dir=3D"ltr"><br></div><div d= ir=3D"ltr">Thank you,</div><div dir=3D"ltr">Patryk</div><div dir=3D"ltr"><br= ><br id=3D"lineBreakAtBeginningOfSignature"><div dir=3D"ltr"><div><span styl= e=3D"background-color: rgba(255, 255, 255, 0);">--</span></div><div id=3D"Ap= pleMailSignature"><span style=3D"background-color: rgba(255, 255, 255, 0);">= Patryk Laurent, Ph.D</span></div><div id=3D"AppleMailSignature" style=3D"dir= ection: inherit;"><span style=3D"background-color: rgba(255, 255, 255, 0);">= San Diego, CA</span></div><div id=3D"AppleMailSignature" style=3D"direction:= inherit;">w: http://pakl.net/</div><div id=3D"AppleMailSignature" style=3D"= direction: inherit;"><br></div></div><blockquote type=3D"cite">On Apr 13, 20= 26, at 07:40, R Frith-Macdonald <[email protected]> wrote:<= br><br></blockquote></div><blockquote type=3D"cite"><div dir=3D"ltr">=EF=BB=BF= <span></span><br><span>On 13/04/2026 04:35, Todd White wrote:</span><br><blo= ckquote type=3D"cite"><span>Hi GNUstep Team,</span><br></blockquote><blockqu= ote type=3D"cite"><span></span><br></blockquote><blockquote type=3D"cite"><s= pan>As an exercise to test out the latest Claude AI capabilities, we recentl= y completed a comprehensive, bottom-up code audit of the GNUstep core stack =E2= =80=94 all seven repositories =E2=80=94 covering libobjc2, libs-base, libs-c= orebase, libs-opal, libs-quartzcore, libs-gui, and libs-back.</span><br></bl= ockquote><span></span><br><span>Thanks for that, it looks like very worthwhi= le work. I can't really comment on other packages, but I'd certainly l= ike to incorporate a lot of that into GNUstep-base. In the areas of th= e base library I can see quite a mixture of stuff, some of which is already k= nown, other parts are quite new. For instance, at one extreme having s= erver verification off by default was a decision from many years back, but o= ne where a change was long overdue, while at the other extreme the check for= cross-thread autorelease pool draining deals with such a perverse situation= (I find it hard to concieve of any way to accidentally do that) that it's q= uite delightful to have a check/fix for it (though classifying it as a criti= cal issue seems silly).</span><br><span></span><br><span>In an ideal world, I= 'd want to see an individual patch for each issue. The patch would con= tain the source code change, the corresponding testcases, and a ChangeLog en= try, but the cloned repos contain only the first of those.</span><br><span><= /span><br><span>Can the LLM take the testcases it has generated and rewrite t= hem as portable (excluding objc2 and/or platform specific features or contro= lling them with preprocessor macros) code as used by the gnustep regression t= esting framework, and add them to the regression tests in each package?</spa= n><br><span></span><br><span>Thanks</span><br><span></span><br><span></span>= <br><span></span><br></div></blockquote></body></html>= --Apple-Mail-CF580FF3-E711-409C-B050-25FBA2C40AA3--