On 2026-01-29 13:36:05 +0800, Kevin J. McCarthy wrote: > On Wed, Jan 28, 2026 at 11:32:57PM -0500, Kurt Hackenberg wrote: > > I'm leery of making reading mbox configurable. This stuff is low-level, > > arcane, and dangerous -- could silently lose messages. Also, making it > > configurable at least doubles the amount of testing needed. > > Yup. But if we want to fix #433, we either have to make the From line > parser much looser for everyone and take that risk, or else make an option.
Without Content-Length, when an "invalid" From line is found, there are 3 possibilities: 1. Accept is as a "From " separator. BTW, for this case, it is also possible to look at the next line: if it does not look like a header, then this choice was incorrect. 2. Assume that it is part of the previous message (except for the first message). 3. Output an error saying that the mailbox is corrupt. For (1) and (2), there is still a possibility for the user to see that something is wrong if the choice was not the right one. #433 is worse than these 3 possibilities as the contents are silently ignored. -- Vincent Lefèvre <[email protected]> - Web: <https://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / Pascaline project (LIP, ENS-Lyon)
