En todos sitios encuentras gente ridicula que quiere hacerse pasar de “guru”, esta conversacion es en el canal de postgresql, un canal dedicado a base de datos, el protagonista es kW_ quien aparentemente tiene un proyecto ultra secreto.
si eres una bestia con el ingles, usa google translate.
<kW_> offtopic: does anybody know a channel where people hang around who work with microsoft compilers+linkers?
<sternocera> kW_: I work with them
<kW_> sternocera: well, do you know how you can tell the microsoft compiler|linker in which order to layout static data in the .data section of a compiled file?
<sternocera> kW_: No, and I don’t know why that matters
<kW_> sternocera: well, it matters in a specific case here… apparently the ordering is not simply the ordering on how the variable definitions for that static data appear
<sternocera> kW_: Do you mean within individual object files?
<kW_> sternocera: well, I mean in the resulting PE file (e.g. .DLL file, .EXE file …)
<sternocera> kW_: I can tell you that the initialisation order of global static data is undefined in C++, so why would the order be defined or predictable?
<kW_> sternocera: well, it is effectively a requirement which popped up from real-live testing, and it is not fun to have it, but it looks like the requirement is not going to go away
<sternocera> kW_: Are you sure it isn’t an imaginary requirement? It sounds like one.
<kW_> sternocera: no, unfortunately not, the consequences are real, I’ve measured them (unfortunately I am not allowed to be more specific)
<sternocera> kW_: You’re funny.
<kW_> sternocera: well, I wish this whole thing was not needed (because I know it is messy and compiler-specific at best), but oddly enough, it is
<sternocera> What are the nature of the consequences?
<RhodiumToad> if you’re trying to force the initialization order, then do it via appropriate methods rather than trying to hack on the physical order of data
<kW_> RhodiumToad: you mean the initialization order would possibly affect the physical order of data (it is the physical order which is needed to be influenced here, not the initialization order)
<RhodiumToad> then ignore me
<sternocera> kW_: Who knows? It probably isn’t documented anywhere, and likely varies between compiler versions. Have fun staring at a debugger looking at an ocean of hex.
<kW_> sternocera: well, this is what I currently do 
Well, I can at least define islands in this ocean and look where they are positioned
<sternocera> kW_: are you aware that you sound like a crazy person?
<kW_> sternocera: well, to be honest, I do not really care about what I sound like. I’m sure however, once you would see the full picture (which I am not allowed to make public, unfortunately), you would understand the requirements.
<sternocera> kW_: sorry, I don’t believe you
<sternocera> kW_: I think it’s more likely that you don’t see the full picture.
<kW_> sternocera: no worries, I perfectly agree with you that it may sound odd (or “unbelievable” ) given not knowing the reason why this is necessary.
<RhodiumToad> if you’re not able to share details of a problem, then it’s probably not a good idea to ask for help solving it
<sternocera> kW_: I actually have the answer to your question, but unfortunately it’s a secret.
<kW_> RhodiumToad: sure… well, I’m not able to share the details of the bigger problem, and I only ask for help for the smaller one…
<sternocera> The problem you have is that you’re asking how to do an incredibly ill-advised and stupid thing, and telling us that you actually have extremely valid but top secret reasons, and that we would totally agree that what you’ve proposed is the way forward if only we were privy to those secret reasons. Ockham’s razor tells me that the simplest explanation is that you’ve decided to do this dumb thing
<sternocera> and have become emotionally attached to the idea
otro tema interesante es que el consumo de recursos de ambos programas es bajisimo.
