Hi! On Fri, Nov 07, 2025 at 04:30:45PM +0900, Hyeonggeun Oh wrote: > Hello haproxy Team. > > I made new patches for > https://www.mail-archive.com/[email protected]&q=subject:%22%5C%5BRFC%5C%5D+dump_all_vars%5C%28%5C%29+implementation+%5C%28issue+%231623%5C%29%22&o=newest > > and it is divided into three patches below. > > The first patch adds settings without affecting existing functionality. > Second patch actually adds the dump_all_vars() function. > Finally, the last part involves Doc-related settings. > > Since this was my first time working on it, it took some time to understand > the structure and make adjustments, and the code may not be perfect yet. > So feedback me please. > > Thank you for always striving to provide excellent haproxy service. > Hyeonggeun Oh (github username Geun-Oh)
So I want to say thank you because your patches are super clean, and it is visible that you've done quite some research in the code to figure how to do that. It's normal that there are comments, but overall there's nothing resulting from big trouble. I'm only unclear about the malloc handling for names given that sometimes they're passed as-is and sometimes allocated, and may also appear as NULL according to the checks. I could be missing something, but I think we should be stricter and consider that starting now, all variables names must be allocated all the time, and if an allocation fails during parsing, we just fail and exit in error, and it will simplify the code. Again, it is possible that I have missed a few points, but at the same time it seems important that all subtleties are easy to grasp when landing on this code. We're polishing 3.3 now (and still facing some difficult but fortunately rare bugs), so I'm not sure whether we can finish to iterate that series before the release. However, we can focus on the infrastructure changes only if you want. I think they're limited enough for a quick merge, and the rest (sample fetch function) is just cosmetic stuff, we can merge it a week or two later and backport it if needed, as it will not have any impact on existing confs since it will just be an addition. Thank you! Willy
